{ "schema_version": "1.0", "generated_at": "2026-10-25T12:38:04.838Z", "verified": "2026-10-24", "base_url": "https://nerdleveltech.com", "count": 21, "open_weights_count": 10, "priced_count": 20, "usage_terms": "Free to read and cite with attribution. Prices, context windows and licence terms are vendor facts copied from the sources in each model entry. Picks are labelled editorial opinions. No benchmark score is published.", "attribution_format": "Source: NerdLevelTech — {url}", "cost_task": { "input_tokens": 50000, "output_tokens": 10000, "note": "USD at the headline API rates for 50,000 input plus 10,000 output tokens. It ignores caching, batch discounts, tier crossings and off-peak pricing; read each model pricing.note." }, "agent": { "mcp": "https://nerdleveltech.com/api/mcp", "a2a": "https://nerdleveltech.com/api/a2a", "tools": [ "search_models", "get_model" ], "index": "https://nerdleveltech.com/models/index.json", "feed": "https://nerdleveltech.com/models-rss.xml", "llms": "https://nerdleveltech.com/llms.txt" }, "picks": [ { "id": "overall", "title": { "en": "Best overall", "ar": "الأفضل عمومًا" }, "qualifier": { "en": "Paid API", "ar": "واجهة مدفوعة" }, "model_slug": "claude-opus-5-5", "model_name": "Claude Opus 5.5", "rationale": { "en": "Anthropic recommends it as the starting point for most workloads, and it backs that up with a 1M context, documented tool use and a $4/$20 price. The default we would ship first.", "ar": "توصي Anthropic به كنقطة بداية لمعظم الأحمال، ويدعم ذلك بسياق مليون رمز واستخدام أدوات موثّق وسعر ٤/٢٠ دولارًا. الخيار الافتراضي الذي نطلقه أولًا." }, "url": "https://nerdleveltech.com/models/claude-opus-5-5", "urls": { "en": "https://nerdleveltech.com/models/claude-opus-5-5", "ar": "https://nerdleveltech.com/ar/models/claude-opus-5-5" } }, { "id": "value", "title": { "en": "Best value", "ar": "أفضل قيمة" }, "qualifier": { "en": "Paid API", "ar": "واجهة مدفوعة" }, "model_slug": "gpt-6-1-sol", "model_name": "GPT-6.1 Sol", "rationale": { "en": "OpenAI positions it as near-Astra performance for complex coding at $2/$10, with the same 1.05M context. The best quality-per-dollar on this list.", "ar": "توصفه OpenAI بأنه أداء قريب من Astra في البرمجة المعقّدة بسعر ٢/١٠ دولارات، مع السياق نفسه (١٫٠٥ مليون رمز). أفضل جودة مقابل الدولار في القائمة." }, "url": "https://nerdleveltech.com/models/gpt-6-1-sol", "urls": { "en": "https://nerdleveltech.com/models/gpt-6-1-sol", "ar": "https://nerdleveltech.com/ar/models/gpt-6-1-sol" } }, { "id": "budget", "title": { "en": "Best budget workhorse", "ar": "أفضل خيار اقتصادي" }, "qualifier": { "en": "Open weights", "ar": "أوزان مفتوحة" }, "model_slug": "deepseek-v4-1-flash", "model_name": "DeepSeek V4.1 Flash", "rationale": { "en": "The cheap agent pick: vision, a 1M-token context window and tool use at about $0.027 per task, half that off-peak. GLM-5.3 Flash is cheaper for text-only work; this is the one we would hand an agent.", "ar": "خيار الوكيل الرخيص: إدخال صوري، وسياق مليون رمز، واستخدام أدوات، بنحو ٠٫٠٢٧ دولارًا للمهمة ونصف ذلك خارج الذروة. GLM-5.3 Flash أرخص للعمل النصي؛ وهذا هو الذي نسلّم إليه وكيلًا." }, "url": "https://nerdleveltech.com/models/deepseek-v4-1-flash", "urls": { "en": "https://nerdleveltech.com/models/deepseek-v4-1-flash", "ar": "https://nerdleveltech.com/ar/models/deepseek-v4-1-flash" } }, { "id": "open-weights", "title": { "en": "Best open weights", "ar": "أفضل أوزان مفتوحة" }, "qualifier": { "en": "Self-host or API", "ar": "استضافة ذاتية أو API" }, "model_slug": "glm-5-3-flash", "model_name": "GLM-5.3 Flash", "rationale": { "en": "The open-weight pick: a dedicated Hugging Face repo, an Ollama library, a 1M context and native image input at $0.15/$0.50 per 1M. The full GLM-5.3 costs about nine times more for the same window.", "ar": "خيار الأوزان المفتوحة: مستودع Hugging Face مخصّص، ومكتبة Ollama، وسياق مليون رمز، وإدخال صوري أصلي، بسعر ٠٫١٥/٠٫٥٠ دولارًا لكل مليون. النسخة الكاملة GLM-5.3 تكلّف نحو تسعة أضعاف للسياق نفسه." }, "url": "https://nerdleveltech.com/models/glm-5-3-flash", "urls": { "en": "https://nerdleveltech.com/models/glm-5-3-flash", "ar": "https://nerdleveltech.com/ar/models/glm-5-3-flash" } } ], "models": [ { "slug": "kimi-k3-1-pro", "name": "Kimi K3.1 Pro", "aliases": [ "kimi-k3.1-pro", "kimi-k3.1-pro-highspeed" ], "maker": "moonshot", "maker_name": "Moonshot AI", "line": "kimi-pro", "status": "current", "released": "2026-10-23", "release_source": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/moonshotai/kimi-k3.1-pro" }, "tagline": { "en": "The 3T-class challenger that undercuts Opus", "ar": "منافس بفئة ٣ ترليون يقلّل سعر Opus" }, "summary": { "en": "Moonshot’s flagship: a 3T-class open-weight model with a 1M context and native visual understanding, positioned for long-horizon coding and knowledge work. The weights are promised by November 4, 2026; the API lists at $2/$10 per 1M tokens.", "ar": "نموذج Moonshot الرائد: أوزان مفتوحة بفئة ٣ ترليون معامل، وسياق مليون رمز، وفهم بصري أصلي، مخصّص للبرمجة الطويلة والأعمال المعرفية. الأوزان موعودة بحلول ٤ نوفمبر ٢٠٢٦؛ والواجهة بسعر ٢/١٠ دولارات لكل مليون رمز." }, "nlt_take": { "en": "The most interesting open-weight challenge to Opus on price: $2/$10 against Opus’s $4/$20, with a 1M window. The caveat is the weights, not the rate.", "ar": "أكثر تحدٍّ مفتوح الأوزان إثارة لـ Opus من ناحية السعر: ٢/١٠ دولارات مقابل ٤/٢٠ لـ Opus، بنافذة مليون رمز. العقبة في الأوزان لا في السعر." }, "watch_out": { "en": "No release date is published in the docs we could read, and the weights land after launch. That is why the card has no Hugging Face link yet.", "ar": "لا تاريخ إصدار منشور في التوثيق الذي اطّلعنا عليه، والأوزان تصدر بعد الإطلاق. لذلك لا تعرض البطاقة رابط Hugging Face بعد." }, "licence": "open-weights", "open_weights_status": "announced", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "coding", "reasoning", "vision", "agentic" ], "best_for": [ "premium-reasoning" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": { "en": "3T-class", "ar": "فئة ٣ ترليون" }, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 2, "output": 10, "cached_input": 0.2 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.2, "urls": { "en": "https://nerdleveltech.com/models/kimi-k3-1-pro", "ar": "https://nerdleveltech.com/ar/models/kimi-k3-1-pro" }, "links": [ { "kind": "maker", "url": "https://www.moonshot.ai/", "label": { "en": "Moonshot AI", "ar": "Moonshot AI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://openrouter.ai/moonshotai/kimi-k3.1-pro", "label": { "en": "OpenRouter listing ($2/$10)", "ar": "قائمة OpenRouter (٢/١٠ دولارات)" }, "supports": [ "price" ], "checked": "2026-10-24", "note": { "en": "Moonshot’s own pricing table is rendered client-side; the USD rate here is the published OpenRouter listing.", "ar": "جدول أسعار Moonshot يُعرض من جانب العميل؛ والسعر الدولاري هنا من قائمة OpenRouter المنشورة." } }, { "kind": "docs", "url": "https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart", "label": { "en": "Kimi K3.1 Pro documentation", "ar": "توثيق Kimi K3.1 Pro" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24", "note": { "en": "Open weights are promised by November 4, 2026; no public Hugging Face repo at launch.", "ar": "الأوزان المفتوحة موعودة بحلول ٤ نوفمبر ٢٠٢٦؛ ولا مستودع Hugging Face عام عند الإطلاق." } }, { "kind": "ollama", "url": "https://ollama.com/library/kimi-k3:cloud", "label": { "en": "Run with Ollama (K3, not K3.1)", "ar": "شغّله عبر Ollama (K3 وليس K3.1)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "Ollama carries Kimi K3 under the cloud tag; K3.1 Pro is not published there yet.", "ar": "يوفّر Ollama نموذج Kimi K3 بوسم cloud؛ وK3.1 Pro لم يُنشر هناك بعد." } } ], "sources": { "pricing": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/moonshotai/kimi-k3.1-pro" }, "specs": { "label": { "en": "Kimi K3.1 Pro documentation", "ar": "توثيق Kimi K3.1 Pro" }, "url": "https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart" } }, "videos": [ { "youtubeId": "5Z_Y7SRd_oo", "language": "en", "title": "Kimi K3.1 Pro Is INSANE! CHEAPER & FASTER Than Claude Opus 5.5?!", "channel": "WorldofAI", "why": { "en": "Tests the cheaper-than-Opus claim, which is the whole pitch.", "ar": "يختبر ادعاء أرخص من Opus، وهو جوهر الطرح." }, "url": "https://www.youtube.com/watch?v=5Z_Y7SRd_oo" }, { "youtubeId": "4MRMLerpcQc", "language": "en", "title": "Kimi K3.1 Pro Is HERE – Does THIS Really Beat Claude & GPT?", "channel": "Bijan Bowen", "why": { "en": "A second opinion on whether it beats the frontier or only undercuts it.", "ar": "رأي ثانٍ: هل يتفوّق على الحدود أم يكتفي بأسعار أقل؟" }, "url": "https://www.youtube.com/watch?v=4MRMLerpcQc" }, { "youtubeId": "9bgM_nJfxZM", "language": "en", "title": "Kimi K3.1 Is the #1 Open Model and It Beat GPT-6 Astra at Coding!", "channel": "Universe of AI", "why": { "en": "The strongest claim in the launch cycle, tested against Astra.", "ar": "أقوى ادعاء في دورة الإطلاق، مختبَرًا مقابل Astra." }, "url": "https://www.youtube.com/watch?v=9bgM_nJfxZM" } ], "related": [ { "slug": "kimi-k3-benchmarks-vs-fable-5-gpt-5-6-grok", "relation": "background" }, { "slug": "kimi-k2-6-300-agent-swarm-open-weight-frontier-coding", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/kimi-k3-1-pro" }, { "slug": "minimax-m3-1-flash", "name": "MiniMax M3.1 Flash", "aliases": [ "MiniMax-M3.1-Flash", "minimax-m3.1-flash:cloud" ], "maker": "minimax", "maker_name": "MiniMax", "line": "minimax-flash", "status": "current", "released": "2026-10-19", "release_source": { "label": { "en": "MiniMax release notes", "ar": "ملاحظات إصدار MiniMax" }, "url": "https://platform.minimaxi.com/docs/release-notes/models" }, "tagline": { "en": "1M context and open weights at a Chinese price", "ar": "سياق مليون رمز وأوزان مفتوحة بسعر صيني" }, "summary": { "en": "MiniMax’s agentic Flash model: a 1M context, native multimodal input, adjustable thinking depth, open weights on Hugging Face, and $0.15/$0.60 per 1M tokens on the global API.", "ar": "نموذج MiniMax الوكيلي من فئة Flash: سياق مليون رمز، وإدخال متعدد الوسائط أصلي، وعمق تفكير قابل للضبط، وأوزان مفتوحة على Hugging Face، وسعر ٠٫١٥/٠٫٦٠ دولارًا لكل مليون رمز على الواجهة العالمية." }, "nlt_take": { "en": "A serious open-weight agent model with a million-token window and a clear USD rate. The remaining question is operational: confirm which MiniMax platform your key belongs to before you put it in a cost model.", "ar": "نموذج وكيل جدي بأوزان مفتوحة ونافذة مليون رمز وسعر دولاري واضح. يبقى السؤال التشغيلي: تحقّق من منصة MiniMax التي ينتمي إليها مفتاحك قبل إدخاله في نموذج التكلفة." }, "watch_out": { "en": "The USD rates are on the global MiniMax platform; the Chinese platform quotes CNY. Use the right one for your key, and check the Ollama tag before pulling.", "ar": "الأسعار الدولارية على منصة MiniMax العالمية؛ والمنصة الصينية تسعّر باليوان. استخدم المنصة الصحيحة لمفتاحك، وتحقّق من وسم Ollama قبل السحب." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "vision", "local" ], "best_for": [ "open-weights", "budget" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.15, "output": 0.6, "cached_input": 0.015 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.0135, "urls": { "en": "https://nerdleveltech.com/models/minimax-m3-1-flash", "ar": "https://nerdleveltech.com/ar/models/minimax-m3-1-flash" }, "links": [ { "kind": "maker", "url": "https://www.minimax.io/", "label": { "en": "MiniMax", "ar": "MiniMax" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://platform.minimax.io/docs/pricing/overview", "label": { "en": "MiniMax pricing (USD)", "ar": "أسعار MiniMax (دولار)" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/minimax-m3.1-flash:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "Ollama has a dedicated MiniMax-M3.1-Flash library; the published tag is cloud.", "ar": "لدى Ollama مكتبة مخصّصة لـ MiniMax-M3.1-Flash؛ والوسم المنشور هو cloud." } } ], "sources": { "pricing": { "label": { "en": "MiniMax pricing (USD)", "ar": "أسعار MiniMax (دولار)" }, "url": "https://platform.minimax.io/docs/pricing/overview" }, "specs": { "label": { "en": "MiniMax model list", "ar": "قائمة نماذج MiniMax" }, "url": "https://platform.minimaxi.com/docs/faq/models" } }, "videos": [ { "youtubeId": "9zVYS00N6mg", "language": "en", "title": "MiniMax M3.1 Flash IS REALLY GOOD! INSANELY FAST & FREE!", "channel": "WorldofAI", "why": { "en": "Tests the speed and free-tier claims on real prompts.", "ar": "يختبر ادعاءات السرعة والطبقة المجانية على طلبات حقيقية." }, "url": "https://www.youtube.com/watch?v=9zVYS00N6mg" }, { "youtubeId": "acPS3TGflXU", "language": "en", "title": "Minimax M3.1 Flash (Fully Tested): Okay, this MODEL is PRETTY GOOD!", "channel": "AICodeKing", "why": { "en": "A second, more measured pass.", "ar": "تجربة ثانية أكثر اتزانًا." }, "url": "https://www.youtube.com/watch?v=acPS3TGflXU" }, { "youtubeId": "lkm-WEHQtNc", "language": "ar", "title": "الموديل المجاني اللي كسر الدنيا… طلع MiniMax!", "channel": "Ahmed Nagdy - أحمد نجدي", "why": { "en": "Arabic coverage of the MiniMax family, focused on the free access angle.", "ar": "تغطية عربية لعائلة MiniMax، بتركيز على الوصول المجاني." }, "url": "https://www.youtube.com/watch?v=lkm-WEHQtNc" } ], "related": [ { "slug": "minimax-m3-open-weight-1m-context-coding-model", "relation": "background" }, { "slug": "minimax-m3-open-weight-coding-sparse-attention", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/minimax-m3-1-flash" }, { "slug": "deepseek-v4-1-flash", "name": "DeepSeek V4.1 Flash", "aliases": [ "deepseek-flash", "DeepSeek-V4.1-Flash-1017", "deepseek-v4.1-flash:1017-cloud" ], "maker": "deepseek", "maker_name": "DeepSeek", "line": "deepseek-flash", "status": "current", "released": "2026-10-17", "release_source": { "label": { "en": "DeepSeek V4.1 release note", "ar": "ملاحظة إصدار DeepSeek V4.1" }, "url": "https://api-docs.deepseek.com/news/news261017" }, "tagline": { "en": "Open weights at a fraction of frontier prices", "ar": "أوزان مفتوحة بجزء بسيط من أسعار النماذج الرائدة" }, "summary": { "en": "DeepSeek’s Flash tier: a 1M-token context window, thinking on by default, vision input and open weights. Peak API pricing is $0.30/$1.20 per 1M tokens; off-peak is half.", "ar": "فئة Flash من DeepSeek: نافذة مليون رمز، وتفكير مُفعّل افتراضيًا، وإدخال صوري، وأوزان مفتوحة. سعر الذروة ٠٫٣٠/١٫٢٠ دولار لكل مليون رمز، ونصف السعر خارج الذروة." }, "nlt_take": { "en": "The best price-to-capability ratio on this list, and the only one you can also run yourself. Use the API for convenience, the weights for privacy.", "ar": "أفضل نسبة سعر إلى قدرة في القائمة، والوحيد الذي يمكنك تشغيله بنفسك أيضًا. استخدم الواجهة للسهولة، والأوزان للخصوصية." }, "watch_out": { "en": "Peak and off-peak prices differ by 2×, so a nightly batch and a daytime agent do not cost the same. Pin the timezone in your cost model.", "ar": "يختلف سعر الذروة عن غيرها بالضعف، فدفعة ليلية ووكيل نهاري لا يكلفان الشيء نفسه. ثبّت المنطقة الزمنية في نموذج التكلفة." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "vision" ], "best_for": [ "open-weights", "budget" ], "picks": [ "budget" ], "context_window": 1000000, "max_input": null, "max_output": 384000, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.3, "output": 1.2, "cached_input": 0.006 }, "note": { "en": "Peak rates. Off-peak is half: $0.15 input / $0.60 output. Peak is 01:00–04:00 and 06:00–10:00 UTC on weekdays.", "ar": "أسعار الذروة. خارج الذروة النصف: ٠٫١٥ دولار إدخال و٠٫٦٠ دولار إخراج. الذروة من ٠١:٠٠ إلى ٠٤:٠٠ ومن ٠٦:٠٠ إلى ١٠:٠٠ بتوقيت UTC أيام العمل." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.027, "urls": { "en": "https://nerdleveltech.com/models/deepseek-v4-1-flash", "ar": "https://nerdleveltech.com/ar/models/deepseek-v4-1-flash" }, "links": [ { "kind": "maker", "url": "https://www.deepseek.com/", "label": { "en": "DeepSeek", "ar": "DeepSeek" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://api-docs.deepseek.com/quick_start/pricing", "label": { "en": "Models and pricing", "ar": "النماذج والأسعار" }, "supports": [ "price", "specs" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/deepseek-v4.1-flash:1017-cloud", "label": { "en": "Run with Ollama (tag 1017-cloud)", "ar": "شغّله عبر Ollama (وسم 1017-cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "The exact V4.1-Flash-1017 build is published as the 1017-cloud tag.", "ar": "البناء الدقيق V4.1-Flash-1017 منشور بوسم 1017-cloud." } } ], "sources": { "pricing": { "label": { "en": "DeepSeek models and pricing", "ar": "نماذج DeepSeek وأسعارها" }, "url": "https://api-docs.deepseek.com/quick_start/pricing" }, "specs": { "label": { "en": "DeepSeek models and pricing", "ar": "نماذج DeepSeek وأسعارها" }, "url": "https://api-docs.deepseek.com/quick_start/pricing" } }, "videos": [ { "youtubeId": "dk1OY5z08_Y", "language": "en", "title": "The New DeepSeek V4.1 Flash Is INSANE – The BEST Open Model Yet?", "channel": "Bijan Bowen", "why": { "en": "A broad test of the open-weight claims.", "ar": "اختبار واسع لادعاءات الأوزان المفتوحة." }, "url": "https://www.youtube.com/watch?v=dk1OY5z08_Y" }, { "youtubeId": "2kEc7ze3xfs", "language": "ar", "title": "DeepSeek V4.1 Flash وصل ... ونهاية V4 Pro رسميًا", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic launch review that also explains the end of V4 Pro.", "ar": "مراجعة عربية للإطلاق تشرح أيضًا نهاية V4 Pro." }, "url": "https://www.youtube.com/watch?v=2kEc7ze3xfs" }, { "youtubeId": "5bFyoew77DA", "language": "ar", "title": "DeepSeek V4 مجاناً بالكامل! بناء تطبيق Full Stack وربطه بقاعدة بيانات", "channel": "كورسات في البرمجة - Korsat X Parmaga", "why": { "en": "Builds a full-stack app in Arabic, which is the practical test of a cheap model.", "ar": "يبني تطبيقًا متكاملًا بالعربية، وهو الاختبار العملي لنموذج رخيص." }, "url": "https://www.youtube.com/watch?v=5bFyoew77DA" } ], "related": [ { "slug": "deepseek-v4-open-source-frontier-million-token-context", "relation": "background" }, { "slug": "deepseek-harness-open-source-agent-loop-api-pricing", "relation": "background" }, { "slug": "china-open-weight-coding-models-cost-war", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/deepseek-v4-1-flash" }, { "slug": "nemotron-3-5-super-vl", "name": "Nemotron 3.5 Super VL", "aliases": [], "maker": "nvidia", "maker_name": "NVIDIA", "line": "nemotron", "status": "current", "released": "2026-10-15", "release_source": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/nvidia/nemotron-3.5-super-vl" }, "tagline": { "en": "NVIDIA’s open vision-language model with a 1M window", "ar": "نموذج NVIDIA المفتوح للرؤية واللغة بسياق مليون رمز" }, "summary": { "en": "NVIDIA’s open vision-language model: a 1,048,576-token context, text and image input, and $0.15/$0.50 per 1M tokens on the hosted API. The public Hugging Face release is the NVFP4 build.", "ar": "نموذج NVIDIA المفتوح للرؤية واللغة: سياق ١٬٠٤٨٬٥٧٦ رمزًا، وإدخال نصي وصوري، وسعر ٠٫١٥/٠٫٥٠ دولارًا لكل مليون رمز على الواجهة المستضافة. والإصدار العام على Hugging Face هو بناء NVFP4." }, "nlt_take": { "en": "The open multimodal option for agent workflows: a 1M window, image input and a real hosted price. It is the cheapest way here to give an agent eyes without closing the weights.", "ar": "الخيار المفتوح متعدد الوسائط لسير عمل الوكلاء: نافذة مليون رمز، وإدخال صوري، وسعر استضافة حقيقي. إنه أرخص طريقة هنا لإعطاء وكيل عيونًا دون إغلاق الأوزان." }, "watch_out": { "en": "Ollama lists the Nemotron family, not this Super VL tag yet. The Hugging Face repo we link is the NVFP4 quantized build, not the full-precision weights.", "ar": "يعرض Ollama عائلة Nemotron، وليس وسم Super VL هذا بعد. ومستودع Hugging Face الذي نربطه هو بناء NVFP4 المكمّم، لا الأوزان كاملة الدقة." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "vision", "agentic", "local", "tool-use" ], "best_for": [ "open-weights", "budget" ], "picks": [], "context_window": 1048576, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": { "en": "120B total / 12B active", "ar": "١٢٠ مليارًا إجمالي / ١٢ مليارًا نشطة" }, "memory": { "label": { "en": "NVFP4", "ar": "NVFP4" }, "note": { "en": "Quantized Hugging Face build", "ar": "بناء Hugging Face مكمّم" } }, "pricing": { "usd_per_million_tokens": { "input": 0.15, "output": 0.5, "cached_input": 0.02 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.0125, "urls": { "en": "https://nerdleveltech.com/models/nemotron-3-5-super-vl", "ar": "https://nerdleveltech.com/ar/models/nemotron-3-5-super-vl" }, "links": [ { "kind": "maker", "url": "https://www.nvidia.com/", "label": { "en": "NVIDIA", "ar": "NVIDIA" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://openrouter.ai/nvidia/nemotron-3.5-super-vl", "label": { "en": "OpenRouter listing ($0.15/$0.50)", "ar": "قائمة OpenRouter (٠٫١٥/٠٫٥٠ دولارًا)" }, "supports": [ "price", "specs" ], "checked": "2026-10-24", "note": { "en": "The USD rate and 1M context here are the published OpenRouter listing.", "ar": "السعر الدولاري وسياق المليون هنا من قائمة OpenRouter المنشورة." } }, { "kind": "weights", "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4", "label": { "en": "Hugging Face weights (NVFP4)", "ar": "الأوزان على Hugging Face (NVFP4)" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/nemotron", "label": { "en": "Run with Ollama (family, not 3.5 Super VL)", "ar": "شغّله عبر Ollama (العائلة، وليس 3.5 Super VL)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "Ollama lists the Nemotron family; the 3.5 Super VL tag is not published there yet.", "ar": "يعرض Ollama عائلة Nemotron؛ ووسم 3.5 Super VL لم يُنشر هناك بعد." } } ], "sources": { "pricing": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/nvidia/nemotron-3.5-super-vl" }, "specs": { "label": { "en": "Hugging Face model card", "ar": "بطاقة النموذج على Hugging Face" }, "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4" } }, "videos": [ { "youtubeId": "P5aLX2nj-WQ", "language": "en", "title": "NVIDIA’s FREE AI: Nemotron 3.5 Super VL… is it worth it? (we tested it on an RTX 5090)", "channel": "Recomienda Tech", "why": { "en": "A local RTX 5090 test, which is how most people will actually run it.", "ar": "اختبار محلي على RTX 5090، وهي الطريقة التي سيشغّله بها معظم الناس." }, "url": "https://www.youtube.com/watch?v=P5aLX2nj-WQ" }, { "youtubeId": "f_rkx4coYOQ", "language": "en", "title": "Build Visual AI Agents With NVIDIA Nemotron 3.5 Super VL", "channel": "NVIDIA Developer", "why": { "en": "The maker’s own walkthrough of the visual-agent use case.", "ar": "شرح الشركة نفسها لحالة استخدام الوكيل البصري." }, "url": "https://www.youtube.com/watch?v=f_rkx4coYOQ" }, { "youtubeId": "7ZbteWZdKQ0", "language": "en", "title": "NVIDIA Nemotron 3.5 Super VL (Full Journey) — 120B Vision-Language Model", "channel": "The Solo Engineer", "why": { "en": "A longer end-to-end look at the 120B model.", "ar": "نظرة أطول من البداية للنهاية على نموذج ١٢٠B." }, "url": "https://www.youtube.com/watch?v=7ZbteWZdKQ0" } ], "related": [ { "slug": "nemo-switchyard-agent-model-routing", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/nemotron-3-5-super-vl" }, { "slug": "gemini-4-argon", "name": "Gemini 4 Argon", "aliases": [], "maker": "google", "maker_name": "Google DeepMind", "line": "gemini-argon", "status": "current", "released": "2026-10-13", "release_source": { "label": { "en": "Gemini API release notes", "ar": "ملاحظات إصدار Gemini API" }, "url": "https://ai.google.dev/gemini-api/docs/changelog" }, "tagline": { "en": "Google’s flagship: 1M context and long decodes", "ar": "نموذج Google الرائد: سياق مليون رمز ومخرجات طويلة" }, "summary": { "en": "Google’s flagship reasoning model, built for software engineering, long-horizon agents and long decodes. One 1,048,576-token window covers input and output; a standard call outputs up to 65,536 tokens, with up to 1M in long-decode mode.", "ar": "نموذج Google الرائد للاستدلال، مصمّم لهندسة البرمجيات والوكلاء الطويلي الأمد والمخرجات الطويلة. نافذة واحدة من ١٬٠٤٨٬٥٧٦ رمزًا تغطي الإدخال والإخراج؛ ويخرج الطلب العادي حتى ٦٥٬٥٣٦ رمزًا، وحتى مليون في وضع long decode." }, "nlt_take": { "en": "The most interesting Google model for coding in a while, and the only one here that can write a whole file set in one response. The introductory price is the reason to try it now.", "ar": "أكثر نموذج من Google إثارة للاهتمام في البرمجة منذ فترة، والوحيد هنا القادر على كتابة مجموعة ملفات كاملة في استجابة واحدة. السعر التمهيدي هو سبب تجربته الآن." }, "watch_out": { "en": "Introductory pricing ($2/$10) ends November 30, 2026; from December 1 it is $4/$20. Budget for the higher number.", "ar": "ينتهي السعر التمهيدي (٢/١٠ دولارات) في ٣٠ نوفمبر ٢٠٢٦؛ وبدءًا من ١ ديسمبر يصبح ٤/٢٠ دولارًا. ضع الميزانية على الرقم الأعلى." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image", "audio", "video", "pdf" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "reasoning", "vision" ], "best_for": [ "long-output" ], "picks": [], "context_window": 1048576, "max_input": null, "max_output": 65536, "context_note": null, "max_output_note": { "en": "Default synchronous output. Up to 1,048,576 tokens with long decode.", "ar": "المخرجات المتزامنة الافتراضية. حتى ١٬٠٤٨٬٥٧٦ رمزًا مع long decode." }, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 2, "output": 10, "cached_input": null }, "note": null, "changes": [ { "date": "2026-12-01", "input": 4, "output": 20, "note": { "en": "Introductory pricing ends November 30, 2026.", "ar": "ينتهي السعر التمهيدي في ٣٠ نوفمبر ٢٠٢٦." } } ] }, "price_label": null, "task_cost_usd": 0.2, "urls": { "en": "https://nerdleveltech.com/models/gemini-4-argon", "ar": "https://nerdleveltech.com/ar/models/gemini-4-argon" }, "links": [ { "kind": "maker", "url": "https://deepmind.google/", "label": { "en": "Google DeepMind", "ar": "Google DeepMind" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://ai.google.dev/gemini-api/docs/pricing", "label": { "en": "Gemini API pricing", "ar": "أسعار Gemini API" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-4-argon", "label": { "en": "Model details", "ar": "تفاصيل النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Gemini API pricing", "ar": "أسعار Gemini API" }, "url": "https://ai.google.dev/gemini-api/docs/pricing" }, "specs": { "label": { "en": "Gemini 4 Argon model details", "ar": "تفاصيل نموذج Gemini 4 Argon" }, "url": "https://ai.google.dev/gemini-api/docs/models/gemini-4-argon" } }, "videos": [ { "youtubeId": "3MMbWDF4xYI", "language": "en", "title": "Gemini 4 Argon Is Here (Fully Tested)", "channel": "Paul J Lipsky", "why": { "en": "A full hands-on pass rather than a launch recap.", "ar": "تجربة عملية كاملة بدلًا من تلخيص الإطلاق." }, "url": "https://www.youtube.com/watch?v=3MMbWDF4xYI" }, { "youtubeId": "xew00eNG7Ao", "language": "ar", "title": "جوجل نزّلت أقوى موديل: Gemini 4 Argon — أربع أشياء لازم تعرفها", "channel": "نايف الأحمدي | تصوير وذكاء اصطناعي", "why": { "en": "Four practical things to know before using Argon, in Arabic.", "ar": "أربعة أمور عملية يجب معرفتها قبل استخدام Argon، بالعربية." }, "url": "https://www.youtube.com/watch?v=xew00eNG7Ao" }, { "youtubeId": "S8xKPMlvPPE", "language": "ar", "title": "Gemini 4 Argon وصل! هل فاق التوقعات.. شاهد النتائج بنفسك", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic results walkthrough for the launch claims.", "ar": "عرض عربي للنتائج مقابل ادعاءات الإطلاق." }, "url": "https://www.youtube.com/watch?v=S8xKPMlvPPE" } ], "related": [ { "slug": "gemini-3-5-flash-benchmarks-pricing-api", "relation": "background" }, { "slug": "ai-agent-token-cost-per-task", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/gemini-4-argon" }, { "slug": "claude-haiku-5-5", "name": "Claude Haiku 5.5", "aliases": [], "maker": "anthropic", "maker_name": "Anthropic", "line": "claude-haiku", "status": "current", "released": "2026-10-07", "release_source": { "label": { "en": "Anthropic release notes", "ar": "ملاحظات إصدار Anthropic" }, "url": "https://docs.claude.com/en/release-notes/api" }, "tagline": { "en": "Cheap, fast, and good enough for the boring 80%", "ar": "سريع ورخيص وكافٍ لنسبة كبيرة من المهام اليومية" }, "summary": { "en": "Anthropic’s cheapest current model, built for high-volume, latency-sensitive work such as classification, extraction and routing. $0.10/$0.50 per 1M tokens up to 100K-token prompts.", "ar": "أرخص نموذج حالي من Anthropic، مصمّم للأعمال كثيفة الحجم والحسّاسة للزمن مثل التصنيف والاستخراج والتوجيه. بسعر ٠٫١٠/٠٫٥٠ دولار لكل مليون رمز حتى ١٠٠ ألف رمز في الطلب." }, "nlt_take": { "en": "The model that makes a multi-step agent affordable: let Haiku classify, extract and route, and call a bigger model only for the step that needs judgement.", "ar": "النموذج الذي يجعل الوكيل متعدد الخطوات ميسور التكلفة: دع Haiku يصنّف ويستخرج ويوجّه، واستدع نموذجًا أكبر فقط للخطوة التي تحتاج حكمًا." }, "watch_out": { "en": "Above 100K-token prompts the price jumps to $0.50/$2.50, which is more than GPT-6 Luna’s base rate. Long-context batch jobs belong on Luna.", "ar": "عند تجاوز ١٠٠ ألف رمز في الطلب يرتفع السعر إلى ٠٫٥٠/٢٫٥٠ دولارًا، وهو أعلى من سعر GPT-6 Luna الأساسي. وظائف الدفعات ذات السياق الطويل أنسب لـ Luna." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "high-volume", "vision", "tool-use" ], "best_for": [ "budget" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": { "en": "Input and output tokens both count toward the 1M window; output is capped at 128K.", "ar": "تُحسب رموز الإدخال والإخراج معًا ضمن نافذة المليون؛ والحد الأقصى للمخرجات ١٢٨ ألف." }, "max_output_note": null, "knowledge_cutoff": "2026-06", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.1, "output": 0.5, "cached_input": null }, "note": { "en": "Up to 100K-token prompts. Above that: $0.50 input / $2.50 output per 1M tokens.", "ar": "حتى ١٠٠ ألف رمز في الطلب. فوق ذلك: ٠٫٥٠ دولار إدخال و٢٫٥٠ دولار إخراج لكل مليون رمز." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.01, "urls": { "en": "https://nerdleveltech.com/models/claude-haiku-5-5", "ar": "https://nerdleveltech.com/ar/models/claude-haiku-5-5" }, "links": [ { "kind": "maker", "url": "https://www.anthropic.com/", "label": { "en": "Anthropic", "ar": "Anthropic" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.claude.com/en/docs/about-claude/pricing", "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.claude.com/en/docs/about-claude/models/overview", "label": { "en": "Model overview", "ar": "نظرة عامة على النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/pricing" }, "specs": { "label": { "en": "Anthropic model overview", "ar": "نظرة عامة على نماذج Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/models/overview" } }, "videos": [ { "youtubeId": "ByzrOG7peAE", "language": "en", "title": "Claude Haiku 5.5 Tested: Is It Better Value Than GPT-6 Luna?", "channel": "Robbert van Empel", "why": { "en": "Compares the two cheapest models here, which is exactly the buy decision.", "ar": "يقارن أرخص نموذجين في القائمة، وهو قرار الشراء نفسه." }, "url": "https://www.youtube.com/watch?v=ByzrOG7peAE" }, { "youtubeId": "h7stDKzzcp0", "language": "ar", "title": "أنثروبيك تقلب الموازين! Claude Haiku 5.5 بسعر مفاجئ وأداء قوي", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic take on the price cut that made Haiku interesting again.", "ar": "نظرة عربية على تخفيض السعر الذي أعاد Haiku إلى الواجهة." }, "url": "https://www.youtube.com/watch?v=h7stDKzzcp0" }, { "youtubeId": "FD7QoktzePE", "language": "ar", "title": "Claude Haiku 5.5 يقلب الموازين! أداء أقوى وتخفيض 75% دفعة واحدة", "channel": "Mohamed Ansary", "why": { "en": "Explains the 75% cut and what it means for high-volume work.", "ar": "يشرح تخفيض ٧٥٪ وما يعنيه للأعمال كثيفة الحجم." }, "url": "https://www.youtube.com/watch?v=FD7QoktzePE" } ], "related": [ { "slug": "claude-sonnet-5-agentic-coding-pricing", "relation": "background" }, { "slug": "ai-agent-token-cost-per-task", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/claude-haiku-5-5" }, { "slug": "mistral-large-4", "name": "Mistral Large 4", "aliases": [ "mistral-large-4", "mistral-large-4-0" ], "maker": "mistral", "maker_name": "Mistral AI", "line": "mistral-large", "status": "current", "released": "2026-10-06", "release_source": { "label": { "en": "Mistral model page", "ar": "صفحة نموذج Mistral" }, "url": "https://docs.mistral.ai/models/mistral-large-4-0" }, "tagline": { "en": "Europe’s 1T challenger", "ar": "التحدّي الأوروبي بترليون معامل" }, "summary": { "en": "Mistral’s 1.05T-parameter mixture-of-experts model with 52B active parameters, a 1M context, text and image input, and an open-weight release promised after the API preview. API pricing is $1.36/$4.18 per 1M tokens.", "ar": "نموذج Mistral بمزيج خبراء ١٫٠٥ ترليون معامل، منها ٥٢ مليارًا نشطة، وسياق مليون رمز، وإدخال نصي وصوري، وأوزان مفتوحة موعودة بعد المعاينة عبر API. سعر الواجهة ١٫٣٦/٤٫١٨ دولارًا لكل مليون رمز." }, "nlt_take": { "en": "The cheapest way to buy near-frontier output from a European vendor, and the only one here with a credible open-weight path at this size. The catch is that the weights are announced, not downloadable yet.", "ar": "أرخص طريقة لشراء مخرجات قريبة من الحدود من شركة أوروبية، والوحيد هنا بمسار أوزان مفتوحة جدّي بهذا الحجم. العقبة أن الأوزان معلنة وليست قابلة للتنزيل بعد." }, "watch_out": { "en": "Open weights are “coming soon”: the Hugging Face repo is not public at launch. Public preview also means the API can change under you.", "ar": "الأوزان المفتوحة «قريبًا»: مستودع Hugging Face غير متاح عند الإطلاق. والمعاينة العامة تعني أن الواجهة قد تتغيّر تحتك." }, "licence": "open-weights", "open_weights_status": "announced", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "reasoning", "vision" ], "best_for": [ "value" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": { "en": "1.05T total / 52B active", "ar": "١٫٠٥ ترليون إجمالي / ٥٢ مليارًا نشطة" }, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 1.36, "output": 4.18, "cached_input": 0.14 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.1098, "urls": { "en": "https://nerdleveltech.com/models/mistral-large-4", "ar": "https://nerdleveltech.com/ar/models/mistral-large-4" }, "links": [ { "kind": "maker", "url": "https://mistral.ai/", "label": { "en": "Mistral AI", "ar": "Mistral AI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.mistral.ai/models/mistral-large-4-0", "label": { "en": "Model card and pricing", "ar": "بطاقة النموذج والأسعار" }, "supports": [ "price", "specs", "licence" ], "checked": "2026-10-24", "note": { "en": "Open weights are announced; the Hugging Face repo is not public yet, so we link the maker page instead of a dead repository.", "ar": "الأوزان المفتوحة معلنة؛ ومستودع Hugging Face ليس عامًا بعد، لذلك نربط صفحة الشركة بدل مستودع فارغ." } }, { "kind": "ollama", "url": "https://ollama.com/library/mistral-large-4:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "The published tag is cloud.", "ar": "الوسم المنشور هو cloud." } } ], "sources": { "pricing": { "label": { "en": "Mistral model page", "ar": "صفحة نموذج Mistral" }, "url": "https://docs.mistral.ai/models/mistral-large-4-0" }, "specs": { "label": { "en": "Mistral model page", "ar": "صفحة نموذج Mistral" }, "url": "https://docs.mistral.ai/models/mistral-large-4-0" } }, "videos": [ { "youtubeId": "OyzP0Rgiz2I", "language": "en", "title": "Mistral Is BACK – Mistral Large 4 First Test (Le Chonk!)", "channel": "Bijan Bowen", "why": { "en": "A first hands-on pass on the 1T model, before the launch gloss settles.", "ar": "أول تجربة عملية على نموذج الترليون، قبل أن يستقرّ بريق الإطلاق." }, "url": "https://www.youtube.com/watch?v=OyzP0Rgiz2I" }, { "youtubeId": "MNOR65DYdpA", "language": "en", "title": "Mistral Large 4 (Fully Tested): Is It Actually The Best Open Model?", "channel": "AICodeKing", "why": { "en": "Tests the “best open model” claim rather than repeating it.", "ar": "يختبر ادعاء «أفضل نموذج مفتوح» بدل ترديده." }, "url": "https://www.youtube.com/watch?v=MNOR65DYdpA" }, { "youtubeId": "hD51W2txi1Y", "language": "en", "title": "Introducing Mistral Large 4, Le Chonk", "channel": "Mistral", "why": { "en": "The maker’s own launch framing, including the 52B-active design.", "ar": "تقديم الشركة نفسه، بما فيه تصميم الـ٥٢ مليار معامل النشطة." }, "url": "https://www.youtube.com/watch?v=hD51W2txi1Y" } ], "related": [ { "slug": "mistral-medium-3-5-open-weight-128b-frontier-coder", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/mistral-large-4" }, { "slug": "gpt-6-1-sol", "name": "GPT-6.1 Sol", "aliases": [], "maker": "openai", "maker_name": "OpenAI", "line": "gpt-sol", "status": "current", "released": "2026-09-29", "release_source": { "label": { "en": "OpenAI changelog", "ar": "سجل تغييرات OpenAI" }, "url": "https://platform.openai.com/docs/changelog" }, "tagline": { "en": "Near-Astra quality at a fifth of the price", "ar": "جودة قريبة من Astra بخُمس السعر" }, "summary": { "en": "OpenAI positions GPT-6.1 Sol as near-Astra performance for complex coding and computer use, at $2/$10 per 1M tokens with the same 1,050,000-token window (922K max input + 128K max output).", "ar": "توصف GPT-6.1 Sol من OpenAI بأنها أداء قريب من Astra في البرمجة المعقّدة واستخدام الحاسوب، بسعر ٢/١٠ دولارات لكل مليون رمز، مع النافذة نفسها (١٬٠٥٠٬٠٠٠ رمز: ٩٢٢ ألف إدخال + ١٢٨ ألف إخراج)." }, "nlt_take": { "en": "The value pick of this list. If you want one model for professional coding without Astra’s bill, start here and move up only when a real task fails.", "ar": "خيار القيمة في هذه القائمة. إذا أردت نموذجًا واحدًا للبرمجة الاحترافية دون فاتورة Astra، فابدأ به وانتقل للأعلى فقط عندما تفشل مهمة حقيقية." }, "watch_out": { "en": "The name is easy to mistype in configs (`gpt-6.1-sol`); a 400 from a typo looks like an outage.", "ar": "الاسم سهل الخطأ في الإعدادات (`gpt-6.1-sol`)؛ وخطأ 400 الناتج عن كتابة خاطئة يبدو كعطل." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "coding", "agentic", "vision", "computer-use" ], "best_for": [ "value" ], "picks": [ "value" ], "context_window": 1050000, "max_input": 922000, "max_output": 128000, "context_note": null, "max_output_note": null, "knowledge_cutoff": "2026-04-30", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 2, "output": 10, "cached_input": 0.1 }, "note": { "en": "Prompts above 272K input tokens: 2× input and cache, 1.5× output for the whole request.", "ar": "الطلبات التي تتجاوز ٢٧٢ ألف رمز إدخالًا: ضعف سعر الإدخال والتخزين المؤقت، و١٫٥ ضعف لسعر الإخراج للطلب كله." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.2, "urls": { "en": "https://nerdleveltech.com/models/gpt-6-1-sol", "ar": "https://nerdleveltech.com/ar/models/gpt-6-1-sol" }, "links": [ { "kind": "maker", "url": "https://openai.com/", "label": { "en": "OpenAI", "ar": "OpenAI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://platform.openai.com/docs/pricing", "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://platform.openai.com/docs/models/gpt-6.1-sol", "label": { "en": "Model documentation", "ar": "توثيق النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "url": "https://platform.openai.com/docs/pricing" }, "specs": { "label": { "en": "GPT-6.1 Sol model page", "ar": "صفحة نموذج GPT-6.1 Sol" }, "url": "https://platform.openai.com/docs/models/gpt-6.1-sol" } }, "videos": [ { "youtubeId": "3t2MhQC5mBk", "language": "en", "title": "GPT-6.1 Sol: What actually makes it useful?", "channel": "Clever Format", "why": { "en": "Focuses on the practical difference from Astra rather than the launch hype.", "ar": "يركّز على الفرق العملي عن Astra بدلًا من ضجيج الإطلاق." }, "url": "https://www.youtube.com/watch?v=3t2MhQC5mBk" }, { "youtubeId": "FFM5Utjnl1Y", "language": "ar", "title": "جرّبت GPT-6.1 Sol", "channel": "Mohamed Kamal", "why": { "en": "A short Arabic hands-on test of the value argument.", "ar": "اختبار عربي قصير لحجة القيمة مقابل السعر." }, "url": "https://www.youtube.com/watch?v=FFM5Utjnl1Y" }, { "youtubeId": "Luabt-FjaeM", "language": "ar", "title": "تجربة GPT-6.1 Sol الجديد.. هل فعلاً بقوة Astra وبخُمس السعر؟!", "channel": "عبداللطيف سعيد Abdellatif Said", "why": { "en": "Directly tests the “near-Astra at a fifth of the price” claim in Arabic.", "ar": "يختبر بالعربية ادعاء «قريب من Astra بخُمس السعر» مباشرة." }, "url": "https://www.youtube.com/watch?v=Luabt-FjaeM" } ], "related": [ { "slug": "openai-dots-always-on-agents-business-guide", "relation": "background" }, { "slug": "openai-agents-api-managed-codex-harness", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/gpt-6-1-sol" }, { "slug": "claude-sonnet-5-5", "name": "Claude Sonnet 5.5", "aliases": [], "maker": "anthropic", "maker_name": "Anthropic", "line": "claude-sonnet", "status": "current", "released": "2026-09-28", "release_source": { "label": { "en": "Anthropic release notes", "ar": "ملاحظات إصدار Anthropic" }, "url": "https://docs.claude.com/en/release-notes/api" }, "tagline": { "en": "The speed/intelligence balance most teams should start with", "ar": "التوازن بين السرعة والذكاء الذي تبدأ به معظم الفرق" }, "summary": { "en": "Anthropic’s middle tier: the same 1M-token context window as Opus, with output capped at 128K, at half the price, with a high default effort and the fastest latency of the current line-up.", "ar": "الفئة الوسطى من Anthropic: نافذة المليون رمز نفسها كـ Opus، مع حد أقصى للمخرجات ١٢٨ ألف، بنصف السعر، بجهد افتراضي مرتفع وأسرع زمن استجابة في التشكيلة الحالية." }, "nlt_take": { "en": "If you are not sure which Claude model to build on, start here and only move up when a failing test says so. The gap to Opus shows up on long agent runs, not on the first prompt.", "ar": "إذا لم تكن متأكدًا من نموذج Claude الذي تبني عليه، فابدأ بهذا ولا تنتقل إلى الأعلى إلا عندما يقول لك اختبار فاشل ذلك. تظهر الفجوة مع Opus في جولات الوكيل الطويلة، لا في أول طلب." }, "watch_out": { "en": "Same price as Sonnet 5 ($2/$10) with better positioning, so there is no reason to pin the older version.", "ar": "السعر نفسه الذي كان لـ Sonnet 5 (٢/١٠ دولارات) مع موقع أفضل، فلا سبب للتمسّك بالإصدار الأقدم." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "vision", "tool-use" ], "best_for": [ "value" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": { "en": "Input and output tokens both count toward the 1M window; output is capped at 128K.", "ar": "تُحسب رموز الإدخال والإخراج معًا ضمن نافذة المليون؛ والحد الأقصى للمخرجات ١٢٨ ألف." }, "max_output_note": null, "knowledge_cutoff": "2026-06", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 2, "output": 10, "cached_input": 0.1 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.2, "urls": { "en": "https://nerdleveltech.com/models/claude-sonnet-5-5", "ar": "https://nerdleveltech.com/ar/models/claude-sonnet-5-5" }, "links": [ { "kind": "maker", "url": "https://www.anthropic.com/", "label": { "en": "Anthropic", "ar": "Anthropic" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.claude.com/en/docs/about-claude/pricing", "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.claude.com/en/docs/about-claude/models/overview", "label": { "en": "Model overview", "ar": "نظرة عامة على النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/pricing" }, "specs": { "label": { "en": "Anthropic model overview", "ar": "نظرة عامة على نماذج Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/models/overview" } }, "videos": [ { "youtubeId": "_31NJMPBby0", "language": "en", "title": "Sonnet 5.5: No-Hype Full Review & Testing", "channel": "Pat Simmons", "why": { "en": "A deliberately unhyped test run, useful before you move a production loop.", "ar": "تجربة متعمّدة بلا مبالغة، مفيدة قبل نقل حلقة إنتاجية." }, "url": "https://www.youtube.com/watch?v=_31NJMPBby0" }, { "youtubeId": "qGzEjREpqqs", "language": "ar", "title": "Claude Sonnet 5.5 رسميًا! أسرع وأرخص وتجربة غير متوقعة", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic launch review focused on the speed and price change, which is the whole point of Sonnet.", "ar": "مراجعة عربية للإطلاق تركّز على تغيّر السرعة والسعر، وهو جوهر Sonnet." }, "url": "https://www.youtube.com/watch?v=qGzEjREpqqs" }, { "youtubeId": "aXntLvp89IE", "language": "ar", "title": "كلود Claude 5.5: Opus ولا Sonnet ولا Haiku؟ وش يناسبك؟", "channel": "نايف الأحمدي | تصوير وذكاء اصطناعي", "why": { "en": "Answers the lineup question in Arabic: which Claude tier fits which job.", "ar": "يجيب بالعربية عن سؤال التشكيلة: أي فئة من Claude تناسب أي عمل." }, "url": "https://www.youtube.com/watch?v=aXntLvp89IE" } ], "related": [ { "slug": "claude-sonnet-5-agentic-coding-pricing", "relation": "background" }, { "slug": "claude-opus-5-5-tool-choice-400-sdks", "relation": "background" }, { "slug": "claude-sonnet-4-6-tool-use-typescript-agentic-loop-tutorial", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/claude-sonnet-5-5" }, { "slug": "claude-opus-5-5", "name": "Claude Opus 5.5", "aliases": [], "maker": "anthropic", "maker_name": "Anthropic", "line": "claude-opus", "status": "current", "released": "2026-09-22", "release_source": { "label": { "en": "Anthropic release notes", "ar": "ملاحظات إصدار Anthropic" }, "url": "https://docs.claude.com/en/release-notes/api" }, "tagline": { "en": "Anthropic's default for long-running agentic coding", "ar": "خيار Anthropic الافتراضي للبرمجة الوكيلية الطويلة" }, "summary": { "en": "Anthropic’s current flagship, positioned for long-running agentic coding and knowledge work. Adaptive thinking is always on, with a medium default effort that you can raise per task.", "ar": "النموذج الرائد الحالي من Anthropic، مخصّص للبرمجة الوكيلية الطويلة والأعمال المعرفية. التفكير التكيّفي مُفعّل دائمًا، بجهد افتراضي متوسط يمكنك رفعه لكل مهمة." }, "nlt_take": { "en": "The safest default if you are building in the Claude ecosystem: documented tool use, a 1M context and less than half of Fable’s price. If your agent loop already works on Sonnet, Opus is the upgrade you make when you can name the test it fails.", "ar": "الخيار الأكثر أمانًا إذا كنت تبني داخل منظومة Claude: استخدام أدوات موثّق، وسياق مليون رمز، وأقل من نصف سعر Fable. إذا كانت حلقة وكيلك تعمل على Sonnet، فإن Opus هي الترقية التي تلجأ إليها عندما تستطيع تحديد الاختبار الذي يفشل فيه." }, "watch_out": { "en": "At $4/$20 it is forty times the output price of Haiku. Route the boring steps to a cheaper model before you make Opus the default for everything.", "ar": "بسعر ٤/٢٠ دولارًا فهو أغلى أربعين مرة من Haiku في المخرجات. وجّه الخطوات الروتينية إلى نموذج أرخص قبل أن تجعل Opus الخيار الافتراضي لكل شيء." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "reasoning", "tool-use", "vision" ], "best_for": [ "overall" ], "picks": [ "overall" ], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": { "en": "Input and output tokens both count toward the 1M window; output is capped at 128K.", "ar": "تُحسب رموز الإدخال والإخراج معًا ضمن نافذة المليون؛ والحد الأقصى للمخرجات ١٢٨ ألف." }, "max_output_note": null, "knowledge_cutoff": "2026-06", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 4, "output": 20, "cached_input": 0.2 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.4, "urls": { "en": "https://nerdleveltech.com/models/claude-opus-5-5", "ar": "https://nerdleveltech.com/ar/models/claude-opus-5-5" }, "links": [ { "kind": "maker", "url": "https://www.anthropic.com/", "label": { "en": "Anthropic", "ar": "Anthropic" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.claude.com/en/docs/about-claude/pricing", "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.claude.com/en/docs/about-claude/models/overview", "label": { "en": "Model overview", "ar": "نظرة عامة على النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/pricing" }, "specs": { "label": { "en": "Anthropic model overview", "ar": "نظرة عامة على نماذج Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/models/overview" } }, "videos": [ { "youtubeId": "ufqJZuh6Y48", "language": "en", "title": "I tested every effort level on Claude Opus 5.5", "channel": "Clearmud", "why": { "en": "Runs the same task at each effort level, which is the setting most teams never tune.", "ar": "ينفّذ المهمة نفسها عند كل مستوى جهد، وهو الإعداد الذي لا تضبطه معظم الفرق." }, "url": "https://www.youtube.com/watch?v=ufqJZuh6Y48" }, { "youtubeId": "gX0L0aFA2xg", "language": "en", "title": "Claude Opus 5.5 is ridiculous", "channel": "AI Search", "why": { "en": "A hands-on look at the agentic coding behaviour, not a spec read-through.", "ar": "نظرة تطبيقية على سلوك البرمجة الوكيلية، لا قراءة مواصفات." }, "url": "https://www.youtube.com/watch?v=gX0L0aFA2xg" }, { "youtubeId": "zqnckVt2VEA", "language": "ar", "title": "شرح Claude Opus 5.5 في 12 دقيقة (بدون مبالغة)", "channel": "Dr. Firas - بالعربية", "why": { "en": "A focused twelve-minute Arabic walkthrough for anyone who wants the whole picture before testing it.", "ar": "شرح عربي مركّز في ١٢ دقيقة لمن يريد الصورة كاملة قبل التجربة." }, "url": "https://www.youtube.com/watch?v=zqnckVt2VEA" }, { "youtubeId": "yPhDxXu2OfY", "language": "ar", "title": "Claude Opus 5.5 | كلود أوبس الجديد أسرع وأقوى وأوفر", "channel": "Mohamed Ansary", "why": { "en": "An Arabic overview of what changed in this release and how it compares with the previous Opus.", "ar": "نظرة عربية على ما تغيّر في هذا الإصدار ومقارنته بـ Opus السابق." }, "url": "https://www.youtube.com/watch?v=yPhDxXu2OfY" } ], "related": [ { "slug": "claude-opus-5-5-tool-choice-400-sdks", "relation": "about" }, { "slug": "claude-inline-tools-beta-tutorial", "relation": "background" }, { "slug": "claude-sonnet-5-agentic-coding-pricing", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/claude-opus-5-5" }, { "slug": "gpt-6-luna", "name": "GPT-6 Luna", "aliases": [], "maker": "openai", "maker_name": "OpenAI", "line": "gpt-luna", "status": "current", "released": "2026-09-22", "release_source": { "label": { "en": "OpenAI changelog", "ar": "سجل تغييرات OpenAI" }, "url": "https://platform.openai.com/docs/changelog" }, "tagline": { "en": "The cheapest model OpenAI ships", "ar": "أرخص نموذج تتيحه OpenAI" }, "summary": { "en": "OpenAI’s high-volume tier at a base rate of $0.10/$0.50 per 1M tokens, with the same 1,050,000-token window (922K max input + 128K max output) as the rest of the GPT-6 family.", "ar": "فئة الحجم الكبير من OpenAI بسعر أساسي ٠٫١٠/٠٫٥٠ دولار لكل مليون رمز، مع النافذة نفسها (١٬٠٥٠٬٠٠٠ رمز: ٩٢٢ ألف إدخال + ١٢٨ ألف إخراج) كبقية عائلة GPT-6." }, "nlt_take": { "en": "The cheapest serious model here, and the one to use for classification, extraction, routing and the twenty steps of an agent that do not need judgement.", "ar": "أرخص نموذج جدّي في القائمة، والأنسب للتصنيف والاستخراج والتوجيه والعشرين خطوة في الوكيل التي لا تحتاج حكمًا." }, "watch_out": { "en": "Cheap tokens are not cheap tasks: a loop that re-sends a growing transcript can cost more than one Astra call. Measure cost per task, not per token.", "ar": "الرموز الرخيصة لا تعني مهام رخيصة: حلقة تعيد إرسال سجل متنامٍ قد تكلف أكثر من استدعاء واحد لـ Astra. قِس التكلفة لكل مهمة، لا لكل رمز." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "high-volume", "vision", "tool-use" ], "best_for": [ "budget" ], "picks": [], "context_window": 1050000, "max_input": 922000, "max_output": 128000, "context_note": null, "max_output_note": null, "knowledge_cutoff": "2026-05-18", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.1, "output": 0.5, "cached_input": 0.01 }, "note": { "en": "Prompts above 272K input tokens: 2× input and cache, 1.5× output for the whole request.", "ar": "الطلبات التي تتجاوز ٢٧٢ ألف رمز إدخالًا: ضعف سعر الإدخال والتخزين المؤقت، و١٫٥ ضعف لسعر الإخراج للطلب كله." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.01, "urls": { "en": "https://nerdleveltech.com/models/gpt-6-luna", "ar": "https://nerdleveltech.com/ar/models/gpt-6-luna" }, "links": [ { "kind": "maker", "url": "https://openai.com/", "label": { "en": "OpenAI", "ar": "OpenAI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://platform.openai.com/docs/pricing", "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://platform.openai.com/docs/models/gpt-6-luna", "label": { "en": "Model documentation", "ar": "توثيق النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "url": "https://platform.openai.com/docs/pricing" }, "specs": { "label": { "en": "GPT-6 Luna model page", "ar": "صفحة نموذج GPT-6 Luna" }, "url": "https://platform.openai.com/docs/models/gpt-6-luna" } }, "videos": [ { "youtubeId": "W9m9S-At4FQ", "language": "en", "title": "GPT-6 Luna First Test – Is OpenAI’s CHEAPEST Model Actually Good?", "channel": "Bijan Bowen", "why": { "en": "Tests the cheapest tier on real prompts instead of assuming it is a toy.", "ar": "يختبر أرخص فئة على طلبات حقيقية بدلًا من افتراض أنها لعبة." }, "url": "https://www.youtube.com/watch?v=W9m9S-At4FQ" }, { "youtubeId": "rmJ20xovr7c", "language": "ar", "title": "GPT-6 Sol و Luna رسميًا! السرعة والسعر تغيّروا تمامًا", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic overview of the Sol and Luna pair, including the new pricing.", "ar": "نظرة عربية على ثنائي Sol و Luna، بما في ذلك الأسعار الجديدة." }, "url": "https://www.youtube.com/watch?v=rmJ20xovr7c" }, { "youtubeId": "j_zGndfjTAk", "language": "ar", "title": "GPT-6 Sol وLuna من OpenAI — كم صار سعر الإدخال والإخراج؟", "channel": "عبدالرحمن AI", "why": { "en": "Breaks down the input/output price change, the reason to use Luna at all.", "ar": "يفصّل تغيّر سعر الإدخال والإخراج، وهو سبب استخدام Luna أصلًا." }, "url": "https://www.youtube.com/watch?v=j_zGndfjTAk" } ], "related": [ { "slug": "ai-agent-token-cost-per-task", "relation": "background" }, { "slug": "ai-decision-models-jev-laya-clef-guide", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/gpt-6-luna" }, { "slug": "grok-4-7", "name": "Grok 4.7", "aliases": [], "maker": "xai", "maker_name": "xAI", "line": "grok", "status": "current", "released": "2026-09-21", "release_source": { "label": { "en": "xAI announcement", "ar": "إعلان xAI" }, "url": "https://x.ai/news/grok-4-7" }, "tagline": { "en": "xAI’s frontier model with a 500K context", "ar": "نموذج xAI الرائد بسياق ٥٠٠ ألف رمز" }, "summary": { "en": "xAI’s frontier model for coding, agentic tasks and knowledge work: a 500,000-token context, text and image input, no stated output limit, and reasoning from low to xhigh.", "ar": "نموذج xAI الرائد للبرمجة والمهام الوكيلية والأعمال المعرفية: سياق ٥٠٠٬٠٠٠ رمز، وإدخال نصي وصوري، وبلا حد معلن للمخرجات، ومستويات تفكير من low إلى xhigh." }, "nlt_take": { "en": "A credible third option for agentic coding with a 500K window and one of the simplest price sheets here. Worth an eval if you want a non-Anthropic, non-OpenAI default.", "ar": "خيار ثالث جدير بالاعتبار للبرمجة الوكيلية بنافذة ٥٠٠ ألف وأبسط جداول الأسعار هنا. يستحق التقييم إذا أردت خيارًا افتراضيًا خارج Anthropic وOpenAI." }, "watch_out": { "en": "Prompts at or above 200K tokens double the price ($4/$12). The 500K window is not a flat-price window.", "ar": "الطلبات عند ٢٠٠ ألف رمز أو أكثر تضاعف السعر (٤/١٢ دولارًا). نافذة ٥٠٠ ألف ليست بسعر ثابت." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "reasoning", "vision" ], "best_for": [ "value" ], "picks": [], "context_window": 500000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": { "en": "xAI states no text output limit.", "ar": "لا تذكر xAI حدًا للمخرجات النصية." }, "knowledge_cutoff": "2026-05", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 2, "output": 6, "cached_input": 0.5 }, "note": { "en": "Prompts under 200K tokens. At 200K or more: $4 input / $12 output for the whole request.", "ar": "الطلبات أقل من ٢٠٠ ألف رمز. عند ٢٠٠ ألف أو أكثر: ٤ دولارات إدخال و١٢ دولارًا إخراج للطلب كله." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.16, "urls": { "en": "https://nerdleveltech.com/models/grok-4-7", "ar": "https://nerdleveltech.com/ar/models/grok-4-7" }, "links": [ { "kind": "maker", "url": "https://x.ai/", "label": { "en": "xAI", "ar": "xAI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.x.ai/docs/models", "label": { "en": "xAI model pricing", "ar": "أسعار نماذج xAI" }, "supports": [ "price", "specs" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.x.ai/developers/grok-4-7", "label": { "en": "Grok 4.7 documentation", "ar": "توثيق Grok 4.7" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "xAI model pricing", "ar": "أسعار نماذج xAI" }, "url": "https://docs.x.ai/docs/models" }, "specs": { "label": { "en": "Grok 4.7 documentation", "ar": "توثيق Grok 4.7" }, "url": "https://docs.x.ai/developers/grok-4-7" } }, "videos": [ { "youtubeId": "AQzt5ZAqEaU", "language": "en", "title": "Is Grok 4.7 Actually Good at Coding? I Tested Everything", "channel": "Arindam Majumder", "why": { "en": "A coding-specific test, the use case this page picks it for.", "ar": "اختبار خاص بالبرمجة، وهي حالة الاستخدام التي تختارها لها الصفحة." }, "url": "https://www.youtube.com/watch?v=AQzt5ZAqEaU" }, { "youtubeId": "1R1qD-95-Ac", "language": "ar", "title": "Grok 4.7: موديل صيني بسعر أمريكي", "channel": "Bashmohandes Mazen - بشمهندس مازن", "why": { "en": "The dissenting Arabic take: is Grok worth its price against the Chinese open models?", "ar": "الرأي العربي المخالف: هل يستحق Grok سعره مقابل النماذج الصينية المفتوحة؟" }, "url": "https://www.youtube.com/watch?v=1R1qD-95-Ac" } ], "related": [ { "slug": "ai-agent-token-cost-per-task", "relation": "background" }, { "slug": "kimi-k3-benchmarks-vs-fable-5-gpt-5-6-grok", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/grok-4-7" }, { "slug": "gpt-6-astra", "name": "GPT-6 Astra", "aliases": [], "maker": "openai", "maker_name": "OpenAI", "line": "gpt-astra", "status": "current", "released": "2026-09-03", "release_source": { "label": { "en": "OpenAI changelog", "ar": "سجل تغييرات OpenAI" }, "url": "https://platform.openai.com/docs/changelog" }, "tagline": { "en": "OpenAI’s most capable model for demanding work", "ar": "أقدر نماذج OpenAI للمهام الصعبة" }, "summary": { "en": "OpenAI’s flagship: a 1,050,000-token window (922K max input + 128K max output), text and image input, and reasoning effort from low to max. Prompts above 272K input tokens are billed at 2× input.", "ar": "النموذج الرائد من OpenAI: نافذة ١٬٠٥٠٬٠٠٠ رمز (٩٢٢ ألف إدخال + ١٢٨ ألف إخراج)، وإدخال نصي وصوري، ومستويات جهد من low إلى max. الطلبات التي تتجاوز ٢٧٢ ألف رمز تُحسب بضعف سعر الإدخال." }, "nlt_take": { "en": "OpenAI’s answer to Opus, at two and a half times the price. Pick it when you are already on OpenAI’s Responses API, need computer use, or your own evals show Astra ahead on your work.", "ar": "رد OpenAI على Opus، بسعر يعادل ضعفيه ونصفًا. اختره إذا كنت تعمل أصلًا على واجهة Responses من OpenAI، أو تحتاج استخدام الحاسوب، أو أظهرت تقييماتك تفوّق Astra في عملك." }, "watch_out": { "en": "The long-context tier is brutal: crossing 272K input tokens reprices the whole request at $20/$75.", "ar": "فئة السياق الطويل مكلفة جدًا: تجاوز ٢٧٢ ألف رمز يعيد تسعير الطلب كله عند ٢٠/٧٥ دولارًا." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "reasoning", "vision", "computer-use" ], "best_for": [ "overall" ], "picks": [], "context_window": 1050000, "max_input": 922000, "max_output": 128000, "context_note": null, "max_output_note": null, "knowledge_cutoff": "2026-04-30", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 10, "output": 50, "cached_input": 1 }, "note": { "en": "Prompts above 272K input tokens: 2× input and cache, 1.5× output for the whole request.", "ar": "الطلبات التي تتجاوز ٢٧٢ ألف رمز إدخالًا: ضعف سعر الإدخال والتخزين المؤقت، و١٫٥ ضعف لسعر الإخراج للطلب كله." }, "changes": [] }, "price_label": null, "task_cost_usd": 1, "urls": { "en": "https://nerdleveltech.com/models/gpt-6-astra", "ar": "https://nerdleveltech.com/ar/models/gpt-6-astra" }, "links": [ { "kind": "maker", "url": "https://openai.com/", "label": { "en": "OpenAI", "ar": "OpenAI" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://platform.openai.com/docs/pricing", "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://platform.openai.com/docs/models/gpt-6-astra", "label": { "en": "Model documentation", "ar": "توثيق النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "OpenAI pricing", "ar": "أسعار OpenAI" }, "url": "https://platform.openai.com/docs/pricing" }, "specs": { "label": { "en": "GPT-6 Astra model page", "ar": "صفحة نموذج GPT-6 Astra" }, "url": "https://platform.openai.com/docs/models/gpt-6-astra" } }, "videos": [ { "youtubeId": "FluKUJyeYD8", "language": "en", "title": "Did OpenAI actually build AGI? GPT-6 Astra first look", "channel": "Fireship", "why": { "en": "A fast, sceptical first look that names what is new and what is marketing.", "ar": "نظرة أولى سريعة ومتشكّكة تسمّي الجديد والتسويق." }, "url": "https://www.youtube.com/watch?v=FluKUJyeYD8" }, { "youtubeId": "FDXAvpo5gL4", "language": "ar", "title": "جربت GPT 6 Astra .. جامد ولا مجرد ارقام ؟", "channel": "مدرسة الذكاء الاصطناعي", "why": { "en": "An Arabic hands-on test that asks whether the launch numbers survive real use.", "ar": "اختبار عربي عملي يسأل: هل تصمد أرقام الإطلاق في الاستخدام الحقيقي؟" }, "url": "https://www.youtube.com/watch?v=FDXAvpo5gL4" }, { "youtubeId": "cOM5tSrF5ug", "language": "ar", "title": "جربت GPT-6 Astra عمليًا… واشتغل مكاني على الكمبيوتر!", "channel": "Heba Ahmed - هبة احمد", "why": { "en": "Focuses on computer use, the capability most teams switch for.", "ar": "يركّز على استخدام الحاسوب، وهي القدرة التي تدفع معظم الفرق للانتقال." }, "url": "https://www.youtube.com/watch?v=cOM5tSrF5ug" } ], "related": [ { "slug": "openai-astra-ten-proofs-lean-verification", "relation": "background" }, { "slug": "openai-agents-api-managed-codex-harness", "relation": "background" }, { "slug": "openai-dots-always-on-agents-business-guide", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/gpt-6-astra" }, { "slug": "gemini-3-8-flash", "name": "Gemini 3.8 Flash", "aliases": [], "maker": "google", "maker_name": "Google DeepMind", "line": "gemini-flash", "status": "current", "released": "2026-09-02", "release_source": { "label": { "en": "Gemini API release notes", "ar": "ملاحظات إصدار Gemini API" }, "url": "https://ai.google.dev/gemini-api/docs/changelog" }, "tagline": { "en": "Google’s fast model for agents and long code", "ar": "نموذج Google السريع للوكلاء والأكواد الطويلة" }, "summary": { "en": "Google’s Flash tier for long-horizon software engineering and agent workflows: 1,048,576-token input, 65,536-token output, text, image, video, audio and PDF input, and a free tier.", "ar": "فئة Flash من Google لهندسة البرمجيات الطويلة وسير عمل الوكلاء: إدخال ١٬٠٤٨٬٥٧٦ رمزًا، وإخراج ٦٥٬٥٣٦ رمزًا، وإدخال نصي وصوري وفيديو وصوتي وPDF، مع طبقة مجانية." }, "nlt_take": { "en": "The cheapest way to put a million-token window behind an agent. The 65K output cap is the catch: it is a reader and a planner, not a whole-repo writer.", "ar": "أرخص طريقة لوضع نافذة مليون رمز خلف وكيل. الحد الأقصى للمخرجات (٦٥ ألف) هو العقبة: إنه قارئ ومخطِّط، لا كاتب لمستودع كامل." }, "watch_out": { "en": "Introductory pricing ends December 31, 2026: $0.75/$3.75 becomes $1.50/$7.50.", "ar": "ينتهي السعر التمهيدي في ٣١ ديسمبر ٢٠٢٦: يصبح ٠٫٧٥/٣٫٧٥ دولارًا ١٫٥٠/٧٫٥٠ دولارًا." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image", "audio", "video", "pdf" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "vision", "tool-use" ], "best_for": [ "long-output" ], "picks": [], "context_window": 1048576, "max_input": null, "max_output": 65536, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.75, "output": 3.75, "cached_input": null }, "note": null, "changes": [ { "date": "2027-01-01", "input": 1.5, "output": 7.5, "note": { "en": "Introductory pricing ends December 31, 2026.", "ar": "ينتهي السعر التمهيدي في ٣١ ديسمبر ٢٠٢٦." } } ] }, "price_label": null, "task_cost_usd": 0.075, "urls": { "en": "https://nerdleveltech.com/models/gemini-3-8-flash", "ar": "https://nerdleveltech.com/ar/models/gemini-3-8-flash" }, "links": [ { "kind": "maker", "url": "https://deepmind.google/", "label": { "en": "Google DeepMind", "ar": "Google DeepMind" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://ai.google.dev/gemini-api/docs/pricing", "label": { "en": "Gemini API pricing", "ar": "أسعار Gemini API" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash", "label": { "en": "Model details", "ar": "تفاصيل النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Gemini API pricing", "ar": "أسعار Gemini API" }, "url": "https://ai.google.dev/gemini-api/docs/pricing" }, "specs": { "label": { "en": "Gemini 3.8 Flash model details", "ar": "تفاصيل نموذج Gemini 3.8 Flash" }, "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash" } }, "videos": [ { "youtubeId": "2uVH2WUYb5E", "language": "en", "title": "GOOGLE IS BACK! (Gemini 3.8 Flash)", "channel": "Matthew Berman", "why": { "en": "The overview video, useful for the model’s positioning in one sitting.", "ar": "فيديو النظرة العامة، مفيد لفهم موقع النموذج في جلسة واحدة." }, "url": "https://www.youtube.com/watch?v=2uVH2WUYb5E" }, { "youtubeId": "NGLcyOSvQIQ", "language": "ar", "title": "تجربة Gemini 3.8 Flash: هل تفوق رسميًا على كل المنافسين في البرمجة؟", "channel": "عبداللطيف سعيد Abdellatif Said", "why": { "en": "An Arabic coding test against the models Flash is usually compared with.", "ar": "اختبار برمجة عربي مقابل النماذج التي تُقارن بها Flash عادةً." }, "url": "https://www.youtube.com/watch?v=NGLcyOSvQIQ" }, { "youtubeId": "brFHoZ0Nn4w", "language": "ar", "title": "جوجل رفعت المستوى! Gemini 3.8 flash يصنع الكثير!", "channel": "Infotech4you - مروي سليمان", "why": { "en": "A short Arabic demo of what the Flash tier can build.", "ar": "عرض عربي قصير لما يمكن لفئة Flash بناؤه." }, "url": "https://www.youtube.com/watch?v=brFHoZ0Nn4w" } ], "related": [ { "slug": "gemini-3-5-flash-benchmarks-pricing-api", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/gemini-3-8-flash" }, { "slug": "muse-spark-1-3", "name": "Muse Spark 1.3", "aliases": [ "muse-spark-1.3", "Muse Spark Open" ], "maker": "meta", "maker_name": "Meta", "line": "muse-spark", "status": "current", "released": "2026-09-02", "release_source": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/meta/muse-spark-1.3" }, "tagline": { "en": "Meta’s agentic coder with a 1M window", "ar": "نموذج Meta الوكيلي بسياق مليون رمز" }, "summary": { "en": "Meta’s agentic and coding model: a 1,048,576-token context, text, image, video, audio and PDF input, and standard pricing of $1.25/$4.25 per 1M tokens. Meta also publishes an open-weight release, Muse Spark Open.", "ar": "نموذج Meta الوكيلي للبرمجة: سياق ١٬٠٤٨٬٥٧٦ رمزًا، وإدخال نصي وصوري وفيديو وصوتي وPDF، بسعر قياسي ١٫٢٥/٤٫٢٥ دولارًا لكل مليون رمز. وتنشر Meta أيضًا إصدارًا مفتوح الأوزان باسم Muse Spark Open." }, "nlt_take": { "en": "The cheapest 1M-context model from a US lab on this list, and the only one with a matching open-weight release. Worth an eval if you want a non-Anthropic, non-OpenAI agent model with image and video input.", "ar": "أرخص نموذج بسياق مليون رمز من مختبر أمريكي في هذه القائمة، والوحيد بإصدار مطابق مفتوح الأوزان. يستحق التقييم إذا أردت نموذجًا وكيليًا خارج Anthropic وOpenAI مع إدخال صوري وفيديو." }, "watch_out": { "en": "Meta does not publish a release date for 1.3 in the docs we could read, and its open release is a separate repo/name (Muse Spark Open), not the hosted endpoint.", "ar": "لا تنشر Meta تاريخ إصدار ١٫٣ في التوثيق الذي اطّلعنا عليه، وإصداره المفتوح مستودع واسم منفصلان (Muse Spark Open)، وليس نقطة النهاية المستضافة." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image", "audio", "video", "pdf" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding", "vision", "local" ], "best_for": [ "value" ], "picks": [], "context_window": 1048576, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 1.25, "output": 4.25, "cached_input": 0.15 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.105, "urls": { "en": "https://nerdleveltech.com/models/muse-spark-1-3", "ar": "https://nerdleveltech.com/ar/models/muse-spark-1-3" }, "links": [ { "kind": "maker", "url": "https://ai.meta.com/", "label": { "en": "Meta", "ar": "Meta" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://dev.meta.ai/docs/pricing-rate-limits", "label": { "en": "Meta Model API pricing", "ar": "أسعار Meta Model API" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://dev.meta.ai/docs/models", "label": { "en": "Muse Spark documentation", "ar": "توثيق Muse Spark" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B", "label": { "en": "Hugging Face weights (Muse Spark Open)", "ar": "الأوزان على Hugging Face (Muse Spark Open)" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24", "note": { "en": "The open release is published as Muse Spark Open; the hosted endpoint is muse-spark-1.3.", "ar": "الإصدار المفتوح منشور باسم Muse Spark Open؛ ونقطة النهاية المستضافة هي muse-spark-1.3." } }, { "kind": "ollama", "url": "https://ollama.com/library/muse-spark-open:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "Ollama carries the open release under the cloud tag, not the hosted endpoint.", "ar": "يوفّر Ollama الإصدار المفتوح بوسم cloud، لا نقطة النهاية المستضافة." } } ], "sources": { "pricing": { "label": { "en": "Meta Model API pricing", "ar": "أسعار Meta Model API" }, "url": "https://dev.meta.ai/docs/pricing-rate-limits" }, "specs": { "label": { "en": "Muse Spark documentation", "ar": "توثيق Muse Spark" }, "url": "https://dev.meta.ai/docs/models" } }, "videos": [ { "youtubeId": "tLlEzZUyGdM", "language": "en", "title": "Meta Muse Spark 1.3 Is HERE – Is THIS a Real Opus Competitor?", "channel": "Bijan Bowen", "why": { "en": "Tests the Opus-competitor framing on real tasks.", "ar": "يختبر طرح «منافس Opus» على مهام حقيقية." }, "url": "https://www.youtube.com/watch?v=tLlEzZUyGdM" }, { "youtubeId": "IDz8XdVuzvM", "language": "ar", "title": "شرح Muse Spark 1.3 أقوى نموذج ذكاء اصطناعي مجاني بديل لـ Claude", "channel": "احمد علام - Ahmed Allam", "why": { "en": "An Arabic walkthrough of the model and its free access.", "ar": "شرح عربي للنموذج وطريقة الوصول المجاني." }, "url": "https://www.youtube.com/watch?v=IDz8XdVuzvM" }, { "youtubeId": "226ktnc3OKg", "language": "ar", "title": "جربت muse spark 1.3 — ما هي افضل اداة ذكاء اصطناعي مجاني جربتها", "channel": "يوميات مطور مواقع", "why": { "en": "A developer’s Arabic hands-on test.", "ar": "تجربة عربية عملية من مطوّر." }, "url": "https://www.youtube.com/watch?v=226ktnc3OKg" } ], "related": [ { "slug": "muse-code-benchmarks-harness-confound", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/muse-spark-1-3" }, { "slug": "claude-fable-5-1", "name": "Claude Fable 5.1", "aliases": [], "maker": "anthropic", "maker_name": "Anthropic", "line": "claude-fable", "status": "current", "released": "2026-09-01", "release_source": { "label": { "en": "Anthropic release notes", "ar": "ملاحظات إصدار Anthropic" }, "url": "https://docs.claude.com/en/release-notes/api" }, "tagline": { "en": "Premium reasoning when Opus is not enough", "ar": "استدلال متقدّم عندما لا يكفي Opus" }, "summary": { "en": "Anthropic’s most demanding-work model: adaptive thinking always on, a 1M-token context window with output capped at 128K, positioned for long-horizon agentic work where Opus at higher effort still falls short.", "ar": "نموذج Anthropic لأصعب المهام: تفكير تكيّفي دائم، ونافذة مليون رمز مع حد أقصى للمخرجات ١٢٨ ألف، مخصّص للأعمال الوكيلية الطويلة التي يظل فيها Opus عند جهد أعلى غير كافٍ." }, "nlt_take": { "en": "A specialist, not a default. At $10/$50 it only pays off when the task is a long reasoning chain where a wrong answer costs more than the tokens.", "ar": "نموذج متخصّص وليس خيارًا افتراضيًا. بسعر ١٠/٥٠ دولارًا لا يستحق التكلفة إلا عندما تكون المهمة سلسلة استدلال طويلة ويكون الخطأ فيها أغلى من الرموز." }, "watch_out": { "en": "Two and a half times Opus’s price for the same 1M context. Benchmark it on your own task before switching a production loop to it.", "ar": "سعر يعادل ضعفَي Opus ونصفًا للسياق نفسه (مليون رمز). قيّمه على مهمتك قبل نقل حلقة إنتاجية إليه." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "reasoning", "agentic", "coding", "vision" ], "best_for": [ "premium-reasoning" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": { "en": "Input and output tokens both count toward the 1M window; output is capped at 128K.", "ar": "تُحسب رموز الإدخال والإخراج معًا ضمن نافذة المليون؛ والحد الأقصى للمخرجات ١٢٨ ألف." }, "max_output_note": null, "knowledge_cutoff": "2026-06", "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 10, "output": 50, "cached_input": 0.25 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 1, "urls": { "en": "https://nerdleveltech.com/models/claude-fable-5-1", "ar": "https://nerdleveltech.com/ar/models/claude-fable-5-1" }, "links": [ { "kind": "maker", "url": "https://www.anthropic.com/", "label": { "en": "Anthropic", "ar": "Anthropic" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.claude.com/en/docs/about-claude/pricing", "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.claude.com/en/docs/about-claude/models/overview", "label": { "en": "Model overview", "ar": "نظرة عامة على النموذج" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "Anthropic pricing", "ar": "أسعار Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/pricing" }, "specs": { "label": { "en": "Anthropic model overview", "ar": "نظرة عامة على نماذج Anthropic" }, "url": "https://docs.claude.com/en/docs/about-claude/models/overview" } }, "videos": [ { "youtubeId": "ROF2Nv_KjOM", "language": "en", "title": "Introducing Claude Fable 5.1", "channel": "Anthropic", "why": { "en": "The vendor’s own framing, before anyone else’s opinion.", "ar": "تقديم الشركة نفسها، قبل أي رأي آخر." }, "url": "https://www.youtube.com/watch?v=ROF2Nv_KjOM" }, { "youtubeId": "9dcdFSFGXSo", "language": "ar", "title": "فيبل 5.1 الميديم أحسن من 5 الهاي وأرخص — إزاي؟", "channel": "Ahmed Nagdy - أحمد نجدي", "why": { "en": "An Arabic test of effort levels: when the cheaper setting is the better one.", "ar": "اختبار عربي لمستويات الجهد: متى يكون الإعداد الأرخص هو الأفضل." }, "url": "https://www.youtube.com/watch?v=9dcdFSFGXSo" }, { "youtubeId": "eupe88VIngs", "language": "ar", "title": "أنثروبيك ضربت الكل بـ Fable 5.1 | أذكى.. بس أرخص؟", "channel": "Heba Ahmed - هبة احمد", "why": { "en": "Weighs the intelligence gain against the price in Arabic.", "ar": "يوازن بين مكسب الذكاء والسعر بالعربية." }, "url": "https://www.youtube.com/watch?v=eupe88VIngs" } ], "related": [ { "slug": "claude-fable-5-mythos-class-model", "relation": "about" }, { "slug": "claude-fable-5-export-controls-relaunch", "relation": "about" }, { "slug": "kimi-k3-benchmarks-vs-fable-5-gpt-5-6-grok", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/claude-fable-5-1" }, { "slug": "glm-5-3-flash", "name": "GLM-5.3 Flash", "aliases": [ "glm-5.3-flash", "glm-5.3-flash:cloud" ], "maker": "zai", "maker_name": "Z.ai", "line": "glm-flash", "status": "current", "released": "2026-08-26", "release_source": { "label": { "en": "Z.ai release notes", "ar": "ملاحظات إصدار Z.ai" }, "url": "https://docs.z.ai/release-notes/new-released" }, "tagline": { "en": "The cheap multimodal GLM", "ar": "GLM الاقتصادي متعدد الوسائط" }, "summary": { "en": "Z.ai’s cheap multimodal coding model: 320B total parameters with 18B active, a 1M context, native image input, and $0.15/$0.50 per 1M tokens. Same post-training family as GLM-5.3, built to be the everyday tier.", "ar": "نموذج Z.ai الاقتصادي متعدد الوسائط للبرمجة: ٣٢٠ مليار معامل منها ١٨ مليارًا نشطة، وسياق مليون رمز، وإدخال صوري أصلي، بسعر ٠٫١٥/٠٫٥٠ دولارًا لكل مليون رمز. من عائلة تدريب GLM-5.3 نفسها، ومصمّم ليكون الفئة اليومية." }, "nlt_take": { "en": "If you want image input and open weights without paying GLM-5.3 prices, this is the lane. It is also the cheapest way here to give an agent eyes on a UI.", "ar": "إذا أردت إدخالًا صوريًا وأوزانًا مفتوحة دون دفع أسعار GLM-5.3، فهذه هي الفئة. وهي أيضًا أرخص طريقة هنا لإعطاء وكيل عيونًا على الواجهة." }, "watch_out": { "en": "The Ollama tag is the GLM-5.3 family page, not a Flash-only build. Check the tag before you pull it.", "ar": "وسم Ollama هو صفحة عائلة GLM-5.3، وليس بناءً خاصًا بـ Flash. تحقّق من الوسم قبل السحب." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "coding", "vision", "local", "high-volume" ], "best_for": [ "open-weights", "budget" ], "picks": [ "open-weights" ], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": { "en": "320B total / 18B active", "ar": "٣٢٠ مليارًا إجمالي / ١٨ مليارًا نشطة" }, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.15, "output": 0.5, "cached_input": 0.03 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.0125, "urls": { "en": "https://nerdleveltech.com/models/glm-5-3-flash", "ar": "https://nerdleveltech.com/ar/models/glm-5-3-flash" }, "links": [ { "kind": "maker", "url": "https://z.ai/", "label": { "en": "Z.ai", "ar": "Z.ai" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.z.ai/guides/overview/pricing", "label": { "en": "Z.ai pricing", "ar": "أسعار Z.ai" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.z.ai/guides/vlm/glm-5.3-flash", "label": { "en": "GLM-5.3 Flash documentation", "ar": "توثيق GLM-5.3 Flash" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/zai-org/GLM-5.3-Flash", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/glm-5.3-flash:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "Ollama has a dedicated GLM-5.3-Flash library; the published tag is cloud.", "ar": "لدى Ollama مكتبة مخصّصة لـ GLM-5.3-Flash؛ والوسم المنشور هو cloud." } } ], "sources": { "pricing": { "label": { "en": "Z.ai pricing", "ar": "أسعار Z.ai" }, "url": "https://docs.z.ai/guides/overview/pricing" }, "specs": { "label": { "en": "GLM-5.3 Flash documentation", "ar": "توثيق GLM-5.3 Flash" }, "url": "https://docs.z.ai/guides/vlm/glm-5.3-flash" } }, "videos": [ { "youtubeId": "VmT7SU81tuM", "language": "en", "title": "GLM 5.3 Flash Is HERE – Is THIS Better Than the FULL GLM 5.3?", "channel": "Bijan Bowen", "why": { "en": "Directly tests the Flash-versus-full question this lane exists for.", "ar": "يختبر مباشرة سؤال Flash مقابل النسخة الكاملة، وهو سبب وجود هذه الفئة." }, "url": "https://www.youtube.com/watch?v=VmT7SU81tuM" }, { "youtubeId": "znoie6-1Bcc", "language": "en", "title": "Is GLM 5.3 Flash The Best Open-Source Model?", "channel": "Hostinger Bytes", "why": { "en": "A second opinion on the open-weight value.", "ar": "رأي ثانٍ في قيمة الأوزان المفتوحة." }, "url": "https://www.youtube.com/watch?v=znoie6-1Bcc" }, { "youtubeId": "S0MN8IU2ccc", "language": "ar", "title": "مراجعة نموذج GLM-5.3.. وليه بلاش تدفع فلوس عشان تجربه؟", "channel": "عبداللطيف سعيد Abdellatif Said", "why": { "en": "Arabic coverage of the GLM-5.3 family, which this Flash tier belongs to.", "ar": "تغطية عربية لعائلة GLM-5.3 التي تنتمي إليها فئة Flash." }, "url": "https://www.youtube.com/watch?v=S0MN8IU2ccc" } ], "related": [ { "slug": "glm-5-2-open-weight-coding-model", "relation": "background" }, { "slug": "chinese-open-weight-coding-llms-three-week-sweep", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/glm-5-3-flash" }, { "slug": "glm-5-3", "name": "GLM-5.3", "aliases": [], "maker": "zai", "maker_name": "Z.ai", "line": "glm", "status": "current", "released": "2026-08-18", "release_source": { "label": { "en": "Z.ai release notes", "ar": "ملاحظات إصدار Z.ai" }, "url": "https://docs.z.ai/release-notes/new-released" }, "tagline": { "en": "The open-weight coder that keeps closing the gap", "ar": "نموذج برمجة مفتوح الأوزان يضيق الفجوة باستمرار" }, "summary": { "en": "Z.ai’s flagship coding model: a 1M-token context window, text-only input, reasoning always on with low/high/max effort. Same base as GLM-5.2, with the gains coming from post-training.", "ar": "نموذج Z.ai الرائد للبرمجة: نافذة مليون رمز، وإدخال نصي فقط، وتفكير مُفعّل دائمًا بمستويات low وhigh وmax. القاعدة نفسها في GLM-5.2، والتحسّن جاء من التدريب اللاحق." }, "nlt_take": { "en": "The open-weight model to try first for coding: a 1M context for a fraction of Opus’s output price, and weights you can host. Text-only input is a fair trade for the price.", "ar": "أول نموذج مفتوح الأوزان تجرّبه للبرمجة: سياق مليون رمز بجزء من سعر مخرجات Opus، وأوزان يمكنك استضافتها. الاقتصار على النص مقايضة عادلة مقابل السعر." }, "watch_out": { "en": "No image input at all. If your workflow screenshots a UI, this is not the model for that step.", "ar": "لا يقبل الصور إطلاقًا. إذا كان سير عملك يعتمد على لقطات شاشة للواجهة، فهذا ليس النموذج المناسب لتلك الخطوة." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text" ], "output": [ "text" ] }, "capabilities": [ "coding", "reasoning", "local" ], "best_for": [ "open-weights" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": 128000, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 1.4, "output": 4.4, "cached_input": 0.26 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.114, "urls": { "en": "https://nerdleveltech.com/models/glm-5-3", "ar": "https://nerdleveltech.com/ar/models/glm-5-3" }, "links": [ { "kind": "maker", "url": "https://z.ai/", "label": { "en": "Z.ai", "ar": "Z.ai" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://docs.z.ai/guides/overview/pricing", "label": { "en": "Z.ai pricing", "ar": "أسعار Z.ai" }, "supports": [ "price" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://docs.z.ai/guides/llm/glm-5.3", "label": { "en": "GLM-5.3 documentation", "ar": "توثيق GLM-5.3" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/zai-org/GLM-5.3", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/glm-5.3:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "The published tag is cloud; check the library page before pulling.", "ar": "الوسم المنشور هو cloud؛ راجع صفحة المكتبة قبل السحب." } } ], "sources": { "pricing": { "label": { "en": "Z.ai pricing", "ar": "أسعار Z.ai" }, "url": "https://docs.z.ai/guides/overview/pricing" }, "specs": { "label": { "en": "GLM-5.3 documentation", "ar": "توثيق GLM-5.3" }, "url": "https://docs.z.ai/guides/llm/glm-5.3" } }, "videos": [ { "youtubeId": "NUJ3ffYR1sQ", "language": "en", "title": "GLM-5.3 Tested: Coding, Security & Creative Coding (Full Review)", "channel": "Fahd Mirza", "why": { "en": "Covers coding and security, the two areas Z.ai claims the biggest gains.", "ar": "يغطي البرمجة والأمان، وهما المجالان اللذان تدّعي Z.ai أكبر تحسّن فيهما." }, "url": "https://www.youtube.com/watch?v=NUJ3ffYR1sQ" }, { "youtubeId": "S0MN8IU2ccc", "language": "ar", "title": "مراجعة نموذج GLM-5.3.. وليه بلاش تدفع فلوس عشان تجربه؟", "channel": "عبداللطيف سعيد Abdellatif Said", "why": { "en": "An Arabic review that asks whether the free/local route is good enough.", "ar": "مراجعة عربية تسأل: هل يكفي المسار المجاني/المحلي؟" }, "url": "https://www.youtube.com/watch?v=S0MN8IU2ccc" }, { "youtubeId": "NY37trQL-0Y", "language": "ar", "title": "GLM 5.3 مذهل! هل أصبح منافسًا حقيقيًا لأقوى نماذج الذكاء الاصطناعي؟", "channel": "Infotech4you - مروي سليمان", "why": { "en": "An Arabic comparison against the proprietary frontier.", "ar": "مقارنة عربية مع النماذج الرائدة المغلقة." }, "url": "https://www.youtube.com/watch?v=NY37trQL-0Y" } ], "related": [ { "slug": "glm-5-2-open-weight-coding-model", "relation": "background" }, { "slug": "glm-5-1-open-source-beats-gpt-coding-benchmarks", "relation": "background" }, { "slug": "chinese-open-weight-coding-llms-three-week-sweep", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/glm-5-3" }, { "slug": "qwen-3-8-27b", "name": "Qwen 3.8 27B", "aliases": [ "qwen3.8:27b", "qwen3.8-27b" ], "maker": "alibaba", "maker_name": "Alibaba (Qwen)", "line": "qwen-local", "status": "current", "released": "2026-08-14", "release_source": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/qwen/qwen3.8-27b" }, "tagline": { "en": "The local 27B that fits on a laptop", "ar": "نموذج ٢٧B المحلي الذي يعمل على حاسوب محمول" }, "summary": { "en": "Alibaba’s open 27B model: a 1M context on the API, text and image input, $0.20/$2.50 per 1M tokens, and an 18GB Ollama build that runs a 256K window.", "ar": "نموذج Alibaba المفتوح بفئة ٢٧B: سياق مليون رمز على الواجهة، وإدخال نصي وصوري، وسعر ٠٫٢٠/٢٫٥٠ دولارًا لكل مليون رمز، وبناء Ollama بحجم ١٨ غيغابايت ونافذة ٢٥٦ ألف رمز." }, "nlt_take": { "en": "The model to try before you pay for a hosted API on privacy-sensitive work. Expect quantization trade-offs and laptop-class throughput, not frontier speed.", "ar": "النموذج الذي تجرّبه قبل الدفع لواجهة مستضافة في الأعمال الحسّاسة للخصوصية. توقّع مقايضات التكميم وسرعة حاسوب محمول، لا سرعة الحدود." }, "watch_out": { "en": "The API context is 1M, but the 18GB Ollama build runs 256K. No release date is published for this version.", "ar": "سياق الواجهة مليون رمز، لكن بناء Ollama بحجم ١٨ غيغابايت يعمل بنافذة ٢٥٦ ألف. ولا تاريخ إصدار منشور لهذا الإصدار." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "local", "vision", "coding", "high-volume" ], "best_for": [ "open-weights", "budget" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": { "en": "27B", "ar": "٢٧B" }, "memory": { "label": { "en": "18GB", "ar": "١٨ غيغابايت" }, "note": { "en": "Ollama default quantized build", "ar": "بناء Ollama الافتراضي بعد التكميم" } }, "pricing": { "usd_per_million_tokens": { "input": 0.2, "output": 2.5, "cached_input": 0.09 }, "note": null, "changes": [] }, "price_label": null, "task_cost_usd": 0.035, "urls": { "en": "https://nerdleveltech.com/models/qwen-3-8-27b", "ar": "https://nerdleveltech.com/ar/models/qwen-3-8-27b" }, "links": [ { "kind": "maker", "url": "https://www.alibabacloud.com/help/en/model-studio/models", "label": { "en": "Alibaba Model Studio", "ar": "Alibaba Model Studio" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://openrouter.ai/qwen/qwen3.8-27b", "label": { "en": "OpenRouter listing ($0.20/$2.50)", "ar": "قائمة OpenRouter (٠٫٢٠/٢٫٥٠ دولارًا)" }, "supports": [ "price", "specs" ], "checked": "2026-10-24", "note": { "en": "Alibaba’s catalog is client-rendered; the USD rates and the 1M API context here are the published OpenRouter listing.", "ar": "دليل Alibaba يُعرض من جانب العميل؛ والسعر الدولاري وسياق المليون هنا من قائمة OpenRouter المنشورة." } }, { "kind": "weights", "url": "https://huggingface.co/Qwen/Qwen3.8-27B", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/qwen3.8:27b", "label": { "en": "Run with Ollama (27b tag, 256K, 18GB)", "ar": "شغّله عبر Ollama (وسم 27b، ٢٥٦ ألف، ١٨ غيغابايت)" }, "supports": [ "local" ], "checked": "2026-10-24" } ], "sources": { "pricing": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/qwen/qwen3.8-27b" }, "specs": { "label": { "en": "OpenRouter listing", "ar": "قائمة OpenRouter" }, "url": "https://openrouter.ai/qwen/qwen3.8-27b" } }, "videos": [ { "youtubeId": "jZ9Mdit5TdI", "language": "en", "title": "QWEN 3.8 27B Local AI Review", "channel": "Digital Spaceport", "why": { "en": "A local-first test, which is the only way this model should be judged.", "ar": "اختبار محلي أولًا، وهي الطريقة الوحيدة للحكم على هذا النموذج." }, "url": "https://www.youtube.com/watch?v=jZ9Mdit5TdI" }, { "youtubeId": "3WbXyUolFA0", "language": "en", "title": "Qwen3.8 27B is something else..", "channel": "Caleb Writes Code", "why": { "en": "Coding-focused impressions from someone running it locally.", "ar": "انطباعات برمجية من شخص يشغّله محليًا." }, "url": "https://www.youtube.com/watch?v=3WbXyUolFA0" }, { "youtubeId": "wm2OVdfr_Oo", "language": "en", "title": "Claude Haiku 5.5 vs Local Qwen 3.8 27B: Is Local AI Still Worth It?", "channel": "Zero to MVP", "why": { "en": "Puts the local model against the cheapest hosted alternative.", "ar": "يضع النموذج المحلي مقابل أرخص بديل مستضاف." }, "url": "https://www.youtube.com/watch?v=wm2OVdfr_Oo" } ], "related": [ { "slug": "build-local-ai-complete-guide-rag-agents-ollama-qwen3", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/qwen-3-8-27b" }, { "slug": "deepseek-v4-pro", "name": "DeepSeek V4 Pro", "aliases": [ "deepseek-v4-pro", "DeepSeek-V4-Pro-0813" ], "maker": "deepseek", "maker_name": "DeepSeek", "line": "deepseek-pro", "status": "current", "released": "2026-08-13", "release_source": { "label": { "en": "DeepSeek V4 Pro GA release note", "ar": "ملاحظة إصدار DeepSeek V4 Pro" }, "url": "https://api-docs.deepseek.com/news/news260813" }, "tagline": { "en": "The Pro tier for hard reasoning, not everyday coding", "ar": "فئة Pro للاستدلال الصعب، لا للبرمجة اليومية" }, "summary": { "en": "DeepSeek’s Pro tier: a 1M-token context window, thinking on by default and open weights. No vision. Peak pricing is $0.66/$1.98 per 1M tokens; off-peak is half.", "ar": "فئة Pro من DeepSeek: نافذة مليون رمز، وتفكير مُفعّل افتراضيًا، وأوزان مفتوحة. بلا إدخال صوري. سعر الذروة ٠٫٦٦/١٫٩٨ دولار لكل مليون رمز، ونصف السعر خارج الذروة." }, "nlt_take": { "en": "Reach for Pro when Flash gets the reasoning wrong, not by default. The price is roughly three times Flash for the same context, and the weights are a real self-host option.", "ar": "استخدم Pro عندما يخطئ Flash في الاستدلال، لا بشكل افتراضي. السعر نحو ثلاثة أضعاف Flash للسياق نفسه، والأوزان خيار استضافة ذاتية حقيقي." }, "watch_out": { "en": "No image input, unlike Flash. If your agent reads screenshots, keep that step on another model.", "ar": "لا يقبل الصور، خلافًا لـ Flash. إذا كان وكيلك يقرأ لقطات شاشة، فأبقِ تلك الخطوة على نموذج آخر." }, "licence": "open-weights", "open_weights_status": "published", "modality": { "input": [ "text" ], "output": [ "text" ] }, "capabilities": [ "reasoning", "coding", "local" ], "best_for": [ "open-weights" ], "picks": [], "context_window": 1000000, "max_input": null, "max_output": 384000, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": { "usd_per_million_tokens": { "input": 0.66, "output": 1.98, "cached_input": 0.044 }, "note": { "en": "Peak rates. Off-peak is half: $0.33 input / $0.99 output.", "ar": "أسعار الذروة. خارج الذروة النصف: ٠٫٣٣ دولار إدخال و٠٫٩٩ دولار إخراج." }, "changes": [] }, "price_label": null, "task_cost_usd": 0.0528, "urls": { "en": "https://nerdleveltech.com/models/deepseek-v4-pro", "ar": "https://nerdleveltech.com/ar/models/deepseek-v4-pro" }, "links": [ { "kind": "maker", "url": "https://www.deepseek.com/", "label": { "en": "DeepSeek", "ar": "DeepSeek" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://api-docs.deepseek.com/quick_start/pricing", "label": { "en": "Models and pricing", "ar": "النماذج والأسعار" }, "supports": [ "price", "specs" ], "checked": "2026-10-24" }, { "kind": "weights", "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813", "label": { "en": "Hugging Face weights", "ar": "الأوزان على Hugging Face" }, "supports": [ "weights", "licence" ], "checked": "2026-10-24" }, { "kind": "ollama", "url": "https://ollama.com/library/deepseek-v4-pro:cloud", "label": { "en": "Run with Ollama (tag cloud)", "ar": "شغّله عبر Ollama (وسم cloud)" }, "supports": [ "local" ], "checked": "2026-10-24", "note": { "en": "The published tag is cloud.", "ar": "الوسم المنشور هو cloud." } } ], "sources": { "pricing": { "label": { "en": "DeepSeek models and pricing", "ar": "نماذج DeepSeek وأسعارها" }, "url": "https://api-docs.deepseek.com/quick_start/pricing" }, "specs": { "label": { "en": "DeepSeek models and pricing", "ar": "نماذج DeepSeek وأسعارها" }, "url": "https://api-docs.deepseek.com/quick_start/pricing" } }, "videos": [ { "youtubeId": "Mu5a8ErscIE", "language": "en", "title": "DeepSeek V4 Pro Is HERE – Is THIS the BEST Open Model Yet?", "channel": "Bijan Bowen", "why": { "en": "A broad test that separates the Pro tier from Flash.", "ar": "اختبار واسع يفصل فئة Pro عن Flash." }, "url": "https://www.youtube.com/watch?v=Mu5a8ErscIE" }, { "youtubeId": "1Aqhg0FVm-w", "language": "en", "title": "Deepseek strikes again", "channel": "AI Search", "why": { "en": "A shorter, sceptical look at what the Pro tier changes.", "ar": "نظرة أقصر وأكثر تشكّكًا في ما تغيّره فئة Pro." }, "url": "https://www.youtube.com/watch?v=1Aqhg0FVm-w" }, { "youtubeId": "vC0mkZBDxG8", "language": "ar", "title": "DeepSeek V4: الجيل الجديد من الذكاء الاصطناعي وصل", "channel": "Mohamed Ansary", "why": { "en": "Arabic background on the V4 family; it predates the Pro tier, so watch it for context.", "ar": "خلفية عربية عن عائلة V4؛ إنها تسبق فئة Pro، فشاهدها للسياق." }, "url": "https://www.youtube.com/watch?v=vC0mkZBDxG8" } ], "related": [ { "slug": "deepseek-v4-open-source-frontier-million-token-context", "relation": "background" }, { "slug": "deepseek-harness-open-source-agent-loop-api-pricing", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/deepseek-v4-pro" }, { "slug": "composer-2-5", "name": "Composer 2.5", "aliases": [ "composer-2.5" ], "maker": "cursor", "maker_name": "Cursor", "line": "cursor-composer", "status": "current", "released": "2026-05-18", "release_source": { "label": { "en": "Cursor changelog", "ar": "سجل تغييرات Cursor" }, "url": "https://cursor.com/changelog" }, "tagline": { "en": "Cursor’s own coding model, built into the editor", "ar": "نموذج Cursor الخاص للبرمجة، مدمج في المحرّر" }, "summary": { "en": "Cursor’s in-house agent model: a 200K context, agentic tool use, a fast interactive tier and a lower cost per token than the frontier APIs. It is part of Cursor’s plans rather than a per-token public API.", "ar": "نموذج Cursor الداخلي للوكلاء: سياق ٢٠٠ ألف رمز، واستخدام أدوات وكيلي، وفئة تفاعلية سريعة، وتكلفة أقل لكل رمز من واجهات الحدود. وهو جزء من خطط Cursor لا واجهة عامة تُسعّر لكل رمز." }, "nlt_take": { "en": "The pragmatic default if your team already lives in Cursor: no API key, no per-token bill, and it is tuned for the editor’s agent loop. It is not a model you can take to another harness.", "ar": "الخيار العملي إذا كان فريقك يعمل أصلًا داخل Cursor: بلا مفتاح API وبلا فاتورة لكل رمز، ومهيّأ لحلقة الوكيل في المحرّر. لكنه ليس نموذجًا يمكن نقله إلى منصة أخرى." }, "watch_out": { "en": "No public per-token API and no Hugging Face or Ollama release. The only way to use it is a Cursor plan, so it cannot be compared on cost like the rest of the table.", "ar": "لا واجهة عامة لكل رمز، ولا إصدار على Hugging Face أو Ollama. الطريقة الوحيدة لاستخدامه هي خطة Cursor، لذا لا يمكن مقارنته بالتكلفة مثل بقية الجدول." }, "licence": "proprietary", "open_weights_status": "not-applicable", "modality": { "input": [ "text", "image" ], "output": [ "text" ] }, "capabilities": [ "agentic", "coding" ], "best_for": [ "value" ], "picks": [], "context_window": 200000, "max_input": null, "max_output": null, "context_note": null, "max_output_note": null, "knowledge_cutoff": null, "params": null, "memory": null, "pricing": null, "price_label": { "en": "Cursor plan", "ar": "ضمن خطة Cursor" }, "task_cost_usd": null, "urls": { "en": "https://nerdleveltech.com/models/composer-2-5", "ar": "https://nerdleveltech.com/ar/models/composer-2-5" }, "links": [ { "kind": "maker", "url": "https://cursor.com/", "label": { "en": "Cursor", "ar": "Cursor" }, "supports": [ "maker" ], "checked": "2026-10-24" }, { "kind": "docs", "url": "https://cursor.com/docs/models/cursor-composer-2-5", "label": { "en": "Composer 2.5 documentation", "ar": "توثيق Composer 2.5" }, "supports": [ "specs", "licence" ], "checked": "2026-10-24" }, { "kind": "pricing", "url": "https://cursor.com/pricing", "label": { "en": "Cursor plans", "ar": "خطط Cursor" }, "supports": [ "price" ], "checked": "2026-10-24", "note": { "en": "Included in Cursor plans; no public per-token API price.", "ar": "مشمول في خطط Cursor؛ ولا سعر API عام لكل رمز." } } ], "sources": { "pricing": { "label": { "en": "Cursor pricing", "ar": "أسعار Cursor" }, "url": "https://cursor.com/pricing" }, "specs": { "label": { "en": "Composer 2.5 documentation", "ar": "توثيق Composer 2.5" }, "url": "https://cursor.com/docs/models/cursor-composer-2-5" } }, "videos": [ { "youtubeId": "hV4nMOmqAVQ", "language": "en", "title": "Cursor Composer 2.5 Is VERY Good – Does THIS Beat GPT & Opus?", "channel": "Bijan Bowen", "why": { "en": "Tests the editor model against the frontier APIs.", "ar": "يختبر نموذج المحرّر مقابل واجهات الحدود." }, "url": "https://www.youtube.com/watch?v=hV4nMOmqAVQ" }, { "youtubeId": "gRRfpCsQG_4", "language": "en", "title": "I Tried Composer 2.5…", "channel": "OrcDev", "why": { "en": "A shorter, developer-perspective trial.", "ar": "تجربة أقصر من منظور مطوّر." }, "url": "https://www.youtube.com/watch?v=gRRfpCsQG_4" }, { "youtubeId": "V1nDUcLju84", "language": "en", "title": "Cursor Composer 2.5: Up to 10x Cheaper, Smarter AI Coding", "channel": "EQOAR", "why": { "en": "Focuses on the cost claim, which is the reason teams pick it.", "ar": "يركّز على ادعاء التكلفة، وهو سبب اختيار الفرق له." }, "url": "https://www.youtube.com/watch?v=V1nDUcLju84" } ], "related": [ { "slug": "spacex-cursor-acquisition-what-it-means-for-developers", "relation": "background" } ], "verified": "2026-10-24", "attribution": "Source: NerdLevelTech — https://nerdleveltech.com/models/composer-2-5" } ] }