Review · Software & AI
Cerebras review
An inference API that serves open models on Cerebras wafer-scale hardware, with published speeds of about 3,000 tokens per second on GPT OSS 120B.
By the DBotMaker research desk · Updated November 16, 2026
- Capabilities (40%)
- 7.6
- Ease of setup (30%)
- 8.6
- Value (30%)
- 8.0
Cerebras is the pick when tokens per second matter more than catalog size. Its pricing page lists GPT OSS 120B at about 3,000 tokens per second for $0.35 per 1M input tokens and $0.75 per 1M output tokens, and Qwen 3.8 27B at about 1,850 tokens per second for $0.99 and $1.49. That makes it roughly three times faster than the speed Groq publishes for the same GPT OSS 120B model, though Groq is cheaper at $0.15 in. The API is OpenAI compatible and supports tool calling, structured outputs and prompt caching, so an agent loop that makes many short calls can finish sooner. New accounts get a one-time $5 credit that expires after 30 days. The limits are real: shared inference carries just two models, context is 131K tokens on paid and 65K on the free trial, and Cerebras describes the Developer tier as meant for development and evaluation, with production capacity and SLAs sold through Dedicated Inference.
Who it is for
Apps where output speed decides the user experience
Price
Pay as you go; GPT OSS 120B costs $0.35 per 1M input tokens and $0.75 per 1M output tokens.
GPT OSS 120B ~3000 tokens/s $0.35/M tokens $0.75/M tokens
Quoted from the vendor pricing page on November 16, 2026.
Key features
- GPT OSS 120B at about 3,000 tokens per second
- Qwen 3.8 27B at about 1,850 tokens per second
- Tool calling, structured outputs and prompt caching
- OpenAI-compatible API
- Dedicated Inference with more model families and production SLAs
Strengths and weaknesses
- Strength: The highest published output speed on this list
- Strength: A one-time $5 credit when you add a payment method
- Strength: Also reachable through OpenRouter, Hugging Face and Vercel
- Weakness: Only two models on shared inference
- Weakness: Cerebras says the Developer tier is not intended for production use
- Weakness: Context of 131K tokens on GPT OSS 120B, 65K on the free trial
Key facts
- GPT OSS 120B: $0.35 in, $0.75 out per 1M tokens
- Qwen 3.8 27B: $0.99 in, $1.49 out per 1M tokens
- Context (free / paid): 65K / 131K on GPT OSS 120B
- Promotional credit: $5, expires after 30 days
Where it ranks
Cerebras is No. 10 in our Best LLM APIs verdict.
Alternatives we scored
Visit cerebras.ai · Questions or corrections: editorial@dbotmaker.io