tools

Cerebras Inference

Cerebras Inference provides API access to fast cloud inference for language and multimodal models.

In inglese

Cerebras Inference lets developers run supported AI models through an OpenAI-compatible API. It is intended for coding, research, voice, automation, agentic workflows, and other applications where response speed matters.

The service offers public endpoints, partner integrations, and dedicated enterprise endpoints. The product page advertises a free trial with $5 in credit; developer usage is pay per token, while enterprise access requires contacting Cerebras. Current pricing tier amounts were not rendered on the pricing page.

Features

OpenAI API compatibility with two code changes
Run supported language and multimodal models through cloud endpoints
Public endpoints for models including GPT OSS 120B and Gemma 4 31B
Dedicated enterprise endpoints for larger model families
Multi-LoRA support for switching adapters per request
Partner access through AWS, OpenRouter, Hugging Face, and Vercel
Free trial with $5 in credit

Use cases

Build responsive coding assistants and coding agents
Run research and knowledge-work agents with lower response latency
Create voice and real-time interactive applications
Automate browser and computer-use workflows
Serve specialized model behavior with multiple LoRA adapters

Pros

Cons

Latest updates

2026-09-22 Temporary increase to Qwen 3.8 27B total token rate limit
The Developer tier total token rate limit for qwen-3.8-27b is temporarily increased from 450K to 750K tokens per minute.
2026-09-03 Gemma 4 31B availability changes on public endpoints
gemma-4-31b is no longer available on Cerebras public endpoints. Gemma 4 31B remains available on Dedicated Endpoints.
2026-09-03 Qwen 3.8 27B available on public endpoints
qwen-3.8-27b is now available on Cerebras public endpoints.
2026-08-28 Flowise integration moved to legacy status
We removed Flowise from the supported integrations directory and navigation. The legacy Flowise guide remains available for existing self-hosted deployments.
2026-07-27 GLM 4.7 reasoning_logprobs default will not ship before deprecation
zai-glm-4.7 will keep its existing default logprobs response shape until deprecation. Requests that explicitly set the X-Cerebras-Version-Patch: 2 header still receive…

Capabilities

Choice of models — “choose the best model for your use case and requirements.” source
API — “OpenAI API compatibility lets developers build on Cerebras with just two code changes.” source
Runs models for you — “The Fastest AI Inference Cloud” source

Security

SOC 2 Type II — “SOC 2 Type 2” source
SOC 2 — “SOC 2 Report” source
GDPR — “LEGAL Data Processing Agreement” source

Pricing

Prices checked: 2026-09-25

Developer Tier — OpenAI GPT OSS 120B

$0.35/M tokens
$0.75/M tokens

~3000 tokens/s

Developer Tier — QWEN Qwen 3.8 27B

$0.99/M tokens
$1.49/M tokens

~1,850 tokens/s

Official website