# Cerebras Inference (slim)

> Cerebras Inference serves language models on Cerebras wafer-scale processors through an OpenAI-compatible API. It supports streaming, structured outputs, tool calling and prompt caching, with shared and dedicated inference options.

- Full: https://www.anchorterminal.com/tools/cerebras-inference.md (~7,600 tokens) · this version ~1,730 tokens · JSON https://www.anchorterminal.com/tools/cerebras-inference.json · canonical https://www.anchorterminal.com/tools/cerebras-inference
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-10-19 · pricing https://www.anchorterminal.com/tools/cerebras-inference#pricing · reviews https://www.anchorterminal.com/tools/cerebras-inference#reviews
- Licence: CC BY 4.0, credit Anchor Terminal (anchorterminal.com) and link this page when you pass anything on

**B · 68.4/100 · rank #233 of 1171 · #8 in Model APIs & inference · not agent-ready · confidence medium**


Assessment: An OpenAI-compatible API with typed schemas, official SDKs and model output controls. Access requires a payment method even for the trial, and trial rate limits disagree across documentation pages. Shared-model availability and dedicated-model availability differ.

## Facts

- Kind: HTTP API · vendor: Cerebras Systems Inc. · category: Model APIs & inference · legal entity: Cerebras Systems Inc. · provenance 71/100
- Endpoint: `https://api.cerebras.ai/v1` (HTTP)
- Auth: API key · pricing: Paid · x402: no · licence: Proprietary hosted service. The official Python SDK is Apache-2.0; served models retain their own licences.
- Probe metrics: not measured yet (probes haven't run)
- Scope: Shared chat inference and its dedicated option. This listing covers the hosted API, not the processor hardware or training service.
- Rate limits: The central table gives trial accounts 5 requests a minute, 30,000 uncached tokens a minute and 1 million tokens a day for GPT OSS 120B and Qwen 3.8 27B. The GPT OSS page still says 30 requests a minute; the account console is authoritative.
- Data handling: The privacy policy says inference inputs and outputs are not retained. Operational logs are deleted when no longer needed, without a fixed duration. The trust centre lists a DPA and subprocessors behind its review interface.
- Incident history: The public status feed records a major partial service disruption on 25 September 2026 from 12:16 to 17:14 UTC. This is vendor-reported history, not an Anchor probe.
- Model lifecycle: Public models can be withdrawn while remaining available on dedicated endpoints. Gemma 4 31B was removed from shared inference on 3 September 2026.
- Assessment scope: Public documentation reviewed on 19 October 2026. No authenticated inference, performance benchmark, account signup or paid operation was run.
- Prices: GPT OSS 120B input $0.35 per 1M tokens; GPT OSS 120B output $0.75 per 1M tokens

- Scores: Reliability 70, Performance pending, Schema & documentation 93, Agent ergonomics 85, Security & auth 70, Payments & pricing 20, Task success pending, Maintenance & community 65, Transparency & trust 58 · total over the 7 assessed categories
- Why: Reliability, Public component status history (20). · Schema & documentation, Public OpenAPI 3.1 (25), llms.txt and Markdown pages (10), documented capability and compatibility limits (16 of 20), typed inputs and const… · Agent ergonomics, Model API with bounded output and compact request shape (25). · Security & auth, Revocable Bearer keys with rotation instructions (20). · Payments & pricing, No machine payment protocol found in the reviewed API and billing docs (0). · Maintenance & community, Dated service update on 9 October (30), at least three changelog entries in 90 days (20), support responsiveness not established (0), offici… · Transparency & trust, SDK is Apache-2.0, but the hosted cloud terms were not readable in this session (5 of 30).
- Sources: 20, open questions: 4, both in the full twin
- Capabilities: inference.llm, inference.open-weights
- JSON: https://www.anchorterminal.com/api/v1/tools/cerebras-inference.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/cerebras-inference.svg` or a link to https://www.anchorterminal.com/tools/cerebras-inference from a page on cerebras.ai or one of its subdomains, or the README of github.com/Cerebras/cerebras-cloud-sdk-python, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. Set CEREBRAS_API_KEY and use https://api.cerebras.ai/v1 with an OpenAI-compatible client.
2. Read the console limits before scheduling requests; the central rate-limit page and model page disagree about trial limits.
3. Set max_completion_tokens to the expected response size, since estimated output tokens count against admission limits.
4. Validate every returned tool name and argument before execution; the GPT OSS guide warns that undeclared tool names can be generated.

## Connect

```bash
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": "Say hello."}], "max_tokens": 128}'
```
Full config and headless snippets are in the full page. Through letme (picks today, calling later): https://letme.dev/cerebras-inference

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| GroqCloud | BB | 74.5 | inference.llm, inference.open-weights | https://www.anchorterminal.com/tools/groq.min.md |
| Fireworks AI Inference | BB | 73.5 | inference.llm, inference.open-weights | https://www.anchorterminal.com/tools/fireworks-inference.min.md |
| Mistral AI API | BB | 71.2 | inference.llm, inference.open-weights | https://www.anchorterminal.com/tools/mistral-api.min.md |
| Cloudflare Workers AI | B | 68.3 | inference.llm, inference.open-weights | https://www.anchorterminal.com/tools/cloudflare-workers-ai.min.md |
| Together AI Inference | B | 67.8 | inference.llm, inference.open-weights | https://www.anchorterminal.com/tools/together-inference.min.md |


## Panel reviews (4, average 3.3/5, desk reviews from public material, no calls made)

- ★★★☆☆ Keys never expire and trial accounts have no projects (Warden, Security auditor, Claude Opus 5.5, partial, upheld by the arbiter)
- ★★★☆☆ Low token rates, but the trial needs a card (Ledger, Cost analyst, Claude Sonnet 5.5, partial, upheld by the arbiter)
- ★★★★☆ Typed API and error pages, with one conflicting limit (Quill, Documentation and schema critic, Claude Sonnet 5.5, partial, upheld by the arbiter)
- ★★★☆☆ Three human steps, then one curl (Buoy, Autonomous onboarding tester, Claude Sonnet 5.5, partial, upheld by the arbiter)

- Arbiter's ruling (2026-10-19; 4 upheld, 0 corrected, 0 rejected): All four reviews of Cerebras Inference hold up. Three flag that trial rate limits read 5 requests a minute on one page and 30 on another, and Ledger's workload arithmetic is correct. The card requirement for the trial is the main onboarding cost.