Skip to content
NerdLevelTech

Alibaba (Qwen)

Qwen 3.8 27B

The local 27B that fits on a laptop

Open weightsReleased: August 14, 2026Self-hostingHigh volumeVerified: October 24, 2026

What it is

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.

NLT take

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.

Before you switch

The API context is 1M, but the 18GB Ollama build runs 256K. No release date is published for this version.

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

$0.035

50K input + 10K output tokens

Input, per 1M
$0.20
Output, per 1M
$2.50
Cached input, per 1M
$0.090

Pricing source: OpenRouter listing

Specs

Maker
Alibaba (Qwen)
Licence
Open weights
Released
August 14, 2026
Context window
1,000,000 tokens
Parameters
27B
Memory (local build)
18GB — Ollama default quantized build
Input / output
Text · Image → Text

Also known as: qwen3.8:27b · qwen3.8-27b

Watch it in action

Hands-on videos from creators we trust. They open on YouTube; we do not host or re-upload them.

From our coverage

Articles about this model or its family. A “Background” tag means we have not covered this exact release yet.

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

Open weights

Run locally