Anthropic
Claude Opus 5.5
Anthropic's default for long-running agentic coding
What it is
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.
NLT take
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.
Before you switch
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.
Cost for a real task
The same task on every model: 50,000 input tokens and 10,000 output tokens, at the base tier.
Task cost
$0.40
50K input + 10K output tokens
- Input, per 1M
- $4.00
- Output, per 1M
- $20.00
- Cached input, per 1M
- $0.20
Pricing source: Anthropic pricing
Specs
- Maker
- Anthropic
- Licence
- Proprietary
- Released
- September 22, 2026
- Context window
- 1,000,000 tokens
- Max output
- 128,000 tokens
- Knowledge cutoff
- June 2026
- Input / output
- Text · Image → Text
Input and output tokens both count toward the 1M window; output is capped at 128K.
Watch it in action
Hands-on videos from creators we trust. They open on YouTube; we do not host or re-upload them.
I tested every effort level on Claude Opus 5.5ClearmudWhy watch: Runs the same task at each effort level, which is the setting most teams never tune.Watch on YouTube(opens YouTube in a new tab)
Claude Opus 5.5 is ridiculousAI SearchWhy watch: A hands-on look at the agentic coding behaviour, not a spec read-through.Watch on YouTube(opens YouTube in a new tab)
From our coverage
Articles about this model or its family. A “Background” tag means we have not covered this exact release yet.
- About this modelSeptember 25, 2026Claude Opus 5.5 tool_choice 400: 3 SDKs Tested (2026)Claude Opus 5.5 rejects forced tool_choice with a 400. I captured what the AI SDK, LangChain and the raw SDK actually send, and built a preflight check.
- BackgroundSeptember 28, 2026Claude Inline Tools Beta: What Actually Ships (2026)Claude's inline tools beta adds tools mid-conversation without busting the prompt cache. I traced exactly when each SDK shipped it, wire format included.
- BackgroundJuly 1, 2026Claude Sonnet 5: Cheaper Agents, Near-Opus Power (2026)Claude Sonnet 5 brings near-Opus 4.8 agentic coding at $2/$10 per million tokens through August. See the pricing, the tokenizer catch, and the system card.
Sources & model links
Every number on this card traces back to one of these links. The maker page proves the positioning, the pricing page proves the rates, the Hugging Face repo proves the weights exist, and the Ollama page proves a local tag.
Official
- AnthropicVerifies: maker
- Anthropic pricingVerifies: price
- Model overviewVerifies: specs · licence