Cerebras API pricing and models
Cerebras publishes per-token prices for 2 text models, as of October 20, 2026. The lowest output rate is $0.75 per million tokens, on GPT OSS 120B, and the highest-scoring model it serves is Qwen3.8-27B, at 73 on our 100-point score.
Cerebras serves models on its wafer-scale chips. Its shared, pay-per-token endpoint lists two models, and the rest of its catalog is sold as dedicated capacity.
Every model Cerebras prices, with its score
Prices are US dollars per million tokens, read from Cerebras pricing page on the date above. The score is our 100-point rating of the model, the same on every host.
| Model | Score | Input /1M | Output /1M | Cached /1M | Context | Listed speed |
|---|---|---|---|---|---|---|
| Qwen 3.8 27Bqwen-3.8-27b | 73 | $0.99 | $1.49 | — | 128K | 1,850 tok/s |
| GPT OSS 120Bgpt-oss-120b | 55 | $0.35 | $0.75 | — | 131K | 3,000 tok/s |
- The self-serve Developer tier is pay as you go. Enterprise capacity is quoted by sales, with no public price.
- Context is shorter on the free trial than on paid accounts. The table shows the paid limits.
We paid for 450 requests on Cerebras
Cerebras hosts neither reference model, so it receives no score. On gpt-oss-120b it was the fastest platform in the run: a 0.35-second median, 1.1-second p95, and no failed requests, on a two-model catalog.
| Model run | Invoiced /1M blended | Median latency | p95 latency | Succeeded |
|---|---|---|---|---|
| gpt-oss-120b | $0.37 | 0.35 s | 1.10 s | 100.0% |
Run on September 1, 2026 with 450 requests per model: short prompts, 20,000-token documents and tool calls. Method and every invoice are on the OpenRouter alternatives audit.
Hosts that sell the same models
Each row counts the scored models both hosts price, and how many of those cost less per output token there than on Cerebras.
Questions
How much does the Cerebras API cost?
Cerebras bills per million tokens, with separate input and output rates for each model. On October 20, 2026 its output rates ran from $0.75 on GPT OSS 120B to $1.49 across 2 paid models. Input is usually cheaper than output, so a long prompt with a short reply costs less than the output rate suggests.
Which models can you run on Cerebras?
2 text models are listed, and 2 of them match a model we score. The highest-scoring is Qwen3.8-27B at 73. The table on this page has every model with its input, output and cached-input rate, the context window, and the date we read the price.
Is Cerebras fast?
In our paid September 1, 2026 run, Cerebras returned a median of 0.35 seconds and a p95 of 1.10 seconds on gpt-oss-120b, over 450 requests. Its docs list 1,850 to 3,000 tokens per second, depending on the model.
What are the alternatives to Cerebras?
OpenRouter lists 2 of the same models and charges less output per token on 2 of them. DeepInfra and Groq also overlap. Price is only part of the choice: speed, rate limits and which model snapshot a host serves differ too.