We may earn a commission if you buy through links on this page. It never changes a score or a ranking. How we make money
Pricing is quote-based.
Cerebras Inference has no published monthly starting price: Developer access is pay-as-you-go per 1 million tokens, while Enterprise pricing is custom. The pricing page lists GPT OSS 120B at $0.35 per million input tokens and $0.75 per million output tokens, and Qwen 3.8 27B at $0.99 input and $1.49 output. A $5 promotional credit requires a valid payment method and expires after 30 days. Developer is not intended for production.
01Every plan, line by line
| Plan | Billed annually | Billed monthly | Built for |
|---|---|---|---|
| Developer | — | Quote | Development, evaluation, and experimentation |
| Enterprise | — | Quote | Production workloads needing tailored capacity |
Add-ons and overages
- PayGo credits; prices not published
- Customized model weights, fine-tuning, and training by agreement; prices not published
02What the plan grid doesn't tell you
Credit expires quickly
The one-time $5 promotional credit requires a valid payment method and expires after 30 days.
Input and output rates differ
The listed rates are per 1 million tokens, and input and output are priced separately. Rates vary by model.
Enterprise pricing is custom
The page does not publish Enterprise monthly or annual pricing; production capacity and terms require a custom quote.
Model work costs extra
Customized weights, fine-tuning, and training are available by agreement, but their prices are not published.
03Which plan should you buy?
| If you are… | Get | Why |
|---|---|---|
| Development and evaluation | Developer | Pay-as-you-go token rates suit experimentation. The plan is not intended for production use. |
| Production workloads | Enterprise | Offers production-ready capacity, higher rate limits, priority, and enhanced latency options; pricing is custom. |
| Teams seeking model customization | Enterprise | Customized weights, fine-tuning, and training are available by agreement, but public pricing is not listed. |
04Is Cerebras Inference worth the money?
Cerebras Inference is easier to price for experimentation than for production: the Developer page gives per-token rates for two models, but not a monthly floor or a complete cost estimate across the model lineup. Enterprise adds production-ready capacity, higher rate limits, priority, and latency options, but requires a custom quote. That flexibility may suit teams with deployment or residency requirements; buyers comparing predictable production spend should treat the missing public quote and hard-to-assess user feedback as real unknowns.
Cerebras Inference pricing FAQ
What is the starting price for Cerebras Inference?
A monthly starting price is not published. Developer usage is pay-as-you-go per 1 million tokens, with listed rates for GPT OSS 120B and Qwen 3.8 27B. Enterprise pricing is custom, so teams should request a quote to estimate production costs.
Does Cerebras Inference have a free plan or trial?
The pricing information does not list a free plan. It does offer a one-time $5 promotional credit with a valid payment method. The credit expires after 30 days, so it is not an ongoing free tier or a published free-trial period.
Can I use the Developer plan for production?
No. The Developer plan is described as being for development, evaluation, and experimentation, and is not intended for production use. Cerebras lists Enterprise as the option with production-ready capacity, higher rate limits, priority, and enhanced latency options.
How much does Cerebras Enterprise cost?
Enterprise pricing is custom, and the pricing page does not publish monthly or annual rates. The plan is for production workloads needing tailored capacity. Customized model weights, fine-tuning, and training are also priced by agreement rather than through a public rate card.