Cerebras
Inference cloud
- Catalog status
- Not auto-checked
- Last successful check
- No successful check yet
- Latest attempt
- No attempt yet
"List checked" means we fetched the provider's model list in the last 3 hours. That shows the list is reachable and current, not that inference works.
Free quota
Trial credit only
$5 trial credit
- Kind
- Trial
- Credit
- $5
- Tokens granted
- Not documented
- Expires after
- 30 days
- Renews every
- Not documented
- Card required
- Yes
- Phone required
- Not documented
- Trains on prompts
- No
- Commercial use
- Not documented
- Eligibility
- New accounts only
- Note
- One-time credit; Cerebras states there is no permanently free tier. Limits are per organization: the Rate Limits page gives 5 RPM, 30K uncached TPM (90K total TPM), and 1M tokens per hour and per day for both Shared Inference models, while the gpt-oss-120b model page lists 30 RPM and 60K input TPM. Free Trial context is 65k tokens, max output 32k. Docs call a payment method, not specifically a card. Tier is described as not intended for production use. Access stops when the credit runs out or expires.
- Source
- Provider page, checked 2026-11-01
| Limits | Per minute | Per day | Per month |
|---|---|---|---|
| Requests | 5 | Not documented | Not documented |
| Tokens | 30,000 | 1,000,000 | Not documented |
Models covered
All modelsHow to call it
- OpenAI-compatible
- Yes, Chat Completions at the base URL
- Base URL
https://api.cerebras.ai/v1
curl "https://api.cerebras.ai/v1/chat/completions" \
-H "Authorization: Bearer $CEREBRAS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "gpt-oss-120b", "messages": [{"role": "user", "content": "Hello"}]}'
Models
| Model ID | Comparison key | Input $/1M | Output $/1M | Price checked |
|---|---|---|---|---|
gpt-oss-120b |
openai/gpt-oss-120b |
$0.35 | $0.75 | models.dev, synced 13h ago |
qwen-3.8-27b |
alibaba/qwen3.8-27b |
$0.99 | $1.49 | models.dev, synced 13h ago |