Cerebras Provider Setup

Cerebras delivers fast, OpenAI-compatible inference running on its own specialized hardware. The plugin includes a fixed catalog of three models, with no dynamic model discovery.

Property	Value
Provider id	`cerebras`
Plugin	official external package (`@openclaw/cerebras-provider`)
Auth env var	`CEREBRAS_API_KEY`
Onboarding flag	`--auth-choice cerebras-api-key`
Direct CLI flag	`--cerebras-api-key <key>`
API	OpenAI-compatible (`openai-completions`)
Base URL	`https://api.cerebras.ai/v1`
Default model	`cerebras/gemma-4-31b`

Install plugin

openclaw plugins install @openclaw/cerebras-provider
openclaw gateway restart

Getting started

Get an API key

Head to the Cerebras Cloud Console to generate an API key.

Run onboarding

openclaw onboard --auth-choice cerebras-api-key

openclaw onboard --non-interactive --accept-risk --skip-health \
  --auth-choice cerebras-api-key \
  --cerebras-api-key "$CEREBRAS_API_KEY"

export CEREBRAS_API_KEY=csk-...

Verify models are available

openclaw models list --provider cerebras

All three models in the static list are shown. When CEREBRAS_API_KEY cannot be resolved, openclaw models status --json flags the absent credential under auth.unusableProfiles.

Non-interactive setup

openclaw onboard --non-interactive --accept-risk --skip-health \
  --mode local \
  --auth-choice cerebras-api-key \
  --cerebras-api-key "$CEREBRAS_API_KEY"

Built-in catalog

Each of the three models supports a 131,072-token context window and caps output at 40,960 tokens.

Model ref	Name	Reasoning	Notes
`cerebras/zai-glm-4.7`	Z.ai GLM 4.7	yes	Scheduled for deprecation August 17, 2026
`cerebras/gpt-oss-120b`	GPT OSS 120B	yes	Production reasoning model
`cerebras/gemma-4-31b`	Gemma 4 31B	yes	Default; preview; text-and-image input

New onboarding follows Cerebras's current Gemma 4 recommendation. Cerebras positions Gemma 4 31B as its reference medium-size model, matching or exceeding GPT OSS in intelligence while adding multimodal agentic capabilities. Since it is a public-preview offering, it can shift or be retired with less advance warning than the production GPT OSS endpoint; OpenClaw setups that already exist keep whichever model they had selected.

Manual config

An API key covers most use cases. To override model metadata or operate in mode: "merge" against the static catalog, apply explicit models.providers.cerebras configuration:

{
  env: { vars: { CEREBRAS_API_KEY: "csk-..." } },
  agents: {
    defaults: {
      model: { primary: "cerebras/gemma-4-31b" },
    },
  },
  models: {
    mode: "merge",
    providers: {
      cerebras: {
        baseUrl: "https://api.cerebras.ai/v1",
        apiKey: "${CEREBRAS_API_KEY}",
        api: "openai-completions",
        models: [
          { id: "zai-glm-4.7", name: "Z.ai GLM 4.7" },
          { id: "gpt-oss-120b", name: "GPT OSS 120B" },
          { id: "gemma-4-31b", name: "Gemma 4 31B" },
        ],
      },
    },
  },
}

Note

When the Gateway runs as a daemon (launchd, systemd, Docker), confirm that CEREBRAS_API_KEY is reachable by that process, for instance in ~/.openclaw/.env or via env.shellEnv. A key set only in an interactive shell will not benefit a managed service unless the environment is imported separately.

Model providers, How to pick providers, model refs, and failover behavior.
Thinking modes, Reasoning effort levels available for Cerebras models.
Configuration reference, Agent defaults and model configuration.
Models FAQ, Auth profiles, switching models, and fixing "no profile" errors.