GitHub radar
Qwen3.8 on Ollama: multimodal in one command
Qwen's new 27B model with native image and video understanding, a 262K-token context, and built-in reasoning is now available locally via Ollama — one command to install.
Qwen3.8-27B is a new multimodal model from Alibaba's Qwen team, released in August 2026. It natively understands text, images, and video within a single model — no extra pipeline needed. The context window is 262,144 tokens (extendable to 1 million), enough to process an entire book or a large codebase in one session. Built-in thinking mode enables step-by-step reasoning and can be toggled per request. On benchmark evaluations it scores 73.0 on Terminal Bench and 61.7 on SWE-bench Pro, per Qwen's official data. Now available on Ollama: ollama run qwen3.8.
Why a vibe-coder should care
Qwen3.8 brings multimodal understanding — analyzing images and video alongside text — to a model that runs entirely on your own hardware. No subscription, no API key, no data sent to someone else's servers. The 262K-token context handles tasks that would hit limits on most cloud APIs.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install Qwen3.8 locally: ollama run qwen3.8 — show me how to send it an image and ask a question about it, and how to enable reasoning mode
A Mac with 32 GB of RAM or a GPU with 16 GB of VRAM required — the model takes up around 18 GB
Open on Hugging Face▌ More finds