Cerebras Provider

Cerebras is exposed as a named preset over SwarmVault's OpenAI-compatible adapter. Cerebras runs inference on purpose-built wafer-scale hardware for fast throughput.

Configuration

{
  "providers": {
    "primary": {
      "type": "cerebras",
      "model": "llama3.1-70b"
    }
  }
}

Options

Field	Default	Description
`model`	--	Model ID (e.g., `llama3.1-70b`, `llama3.1-8b`)
`apiKeyEnv`	`"CEREBRAS_API_KEY"`	Environment variable for API key
`baseUrl`	`https://api.cerebras.ai/v1`	API base URL
`apiStyle`	`"chat"`	API style (`chat` for Chat Completions)
`capabilities`	--	Override auto-detected capabilities

Environment Variable

export CEREBRAS_API_KEY=csk-...

Notes

Speed: Cerebras is optimized for high-throughput inference on their custom hardware. Like Groq, it works well as a fast provider for query or lint tasks.
Models: Cerebras currently hosts Llama variants. Check the Cerebras API docs for the current model list.