Whether you’re building agents, refactoring legacy code, or exploring new codebases, you can now work at the speed of thought using the latest open models like Qwen3 Coder and Qwen3-235B — powered by Cerebras’s wafer-scale inference.

​
Set Up Cerebras in RooCode

1

Get a free Cerebras API key

Get an API key here. Step 1
2

Install RooCode

Head to RooCode and click Install RooCode. Step 2
3

Configure your Cerebras API key

  • Click on the settings tab: Step 3
  • Select Cerebras as a provider: Step 3-1
  • Paste in your API key: Step 3-2
4

Choose a Cerebras-backed model

In the model selector, pick Qwen-3-32B, Qwen-3-235B-a22b, Qwen-3-235B-a22b-instruct-2507 or Llama3.3-70B to tap into lightning-fast inference. Step 4
5

Start coding — but faster

Use RooCode as usual: refactor, generate, build agents, or explore codebases — now with near-zero latency
6

Need more speed or scale?

Upgrade with Cerebras credits for 24/7 access to premium models like Qwen3-235B.