Community model

Qwen 3.5 9B on Braise

Use Qwen 3.5 9B through the Braise local API without owning a GPU, or host it on your own machine and earn Compute Credits for each completed job.

Theoretical model capacity

Context window: 262,144 tokens
Maximum response: 81,920 tokens

Published limits of the model itself.

Use it

In the Braise app, point your AI tool at the local API and request this model by its id. You don’t need a GPU, a runtime or a local copy of the model.

Model id

braise/qwen3.5:9b

curl $BRAISE_API/v1/chat/completions \
  -H "Authorization: Bearer $BRAISE_KEY" \
  -d '{"model": "braise/qwen3.5:9b", "messages": [...]}'

Availability depends on community machines online for this model. If none is available, the request fails with a capacity error and isn’t charged.

I want to use Braise

Host it

These runtime and platform combinations can serve this model on the network. The provider guide has the exact commands and downloads.

Runtime	Platforms	Download size
llama.cpp	Mac · Apple Silicon	~5.7 GB
Ollama	Mac · Apple Silicon	~6.6 GB
oMLX	Mac · Apple Silicon	~6.3 GB

I want to share my machine Read the provider guide

The machine that runs a request can read it and its response. Use public or made-up data, never confidential information.