Cerebras API Pricing (2026)
Cerebras delivers the world's fastest LLM inference using custom wafer-scale chips, achieving up to 3,000 tokens/second — roughly 20x faster than GPU-based providers — at competitive prices. Founded in 2016 and headquartered in Sunnyvale, CA, Cerebras offers 3 active models through their API. All pricing data below is sourced directly from the Cerebras pricing page and is updated regularly.
Prices verified Mar 10, 2026
Cerebras Model Pricing — All Models
All prices are in USD per 1 million tokens ($/M tokens). Input tokens are the text you send to the model; output tokens are the generated response.
| Model | Tier | Input | Output | Cached Input | Batch Input | Batch Output |
|---|---|---|---|---|---|---|
Llama 3.1 8B (Cerebras) Meta's Llama 3.1 8B on Cerebras wafer-scale hardware — achieving ~2,200 tokens/sec inference speed. The fastest available serving of an 8B model at an extremely low price. | Budget | $0.10/M | $0.10/M | — | — | — |
GPT OSS 120B (Cerebras) OpenAI's open-weight GPT OSS 120B on Cerebras hardware, achieving ~3,000 tokens/sec. Production-ready large open-source model at exceptional inference speeds. | Mid | $0.35/M | $0.75/M | — | — | — |
Qwen3 235B Instruct (Cerebras) Alibaba's Qwen3 235B model on Cerebras, delivering ~1,400 tokens/sec. A preview of one of the largest open-weight models available at ultra-fast inference speeds. | Premium | $0.60/M | $1.20/M | — | — | — |
Prices last verified: 2026-03-10. Always confirm on the official pricing page before production use.
Cerebras Model Capabilities
Key technical capabilities across all Cerebras models. Use this table to identify which models support the features your application requires.
| Model | Context | Max Output | Vision | Functions | JSON Mode | Caching | Batch API | Reasoning |
|---|---|---|---|---|---|---|---|---|
| Llama 3.1 8B (Cerebras) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| GPT OSS 120B (Cerebras) | 128,000 | 16,384 | — | ✓ | ✓ | — | — | — |
| Qwen3 235B Instruct (Cerebras) | 128,000 | 32,768 | — | ✓ | ✓ | — | — | ✓ |
When to Choose Cerebras
Cost-sensitive workloads
Use Llama 3.1 8B (Cerebras) — Cerebras's most affordable model at $0.10/M input / $0.10/M output. Estimated cost at 100K requests/month: $15.00.
Complex or high-accuracy tasks
Use Qwen3 235B Instruct (Cerebras) — Cerebras's most capable model (premium tier) with a 128,000-token context window. Priced at $0.60/M input / $1.20/M output.
High-volume pipelines
Cerebras models are well-suited for high-throughput workloads. Consider batch API (where supported) for an additional 50% cost reduction on asynchronous pipelines. Check the pricing table above for batch pricing availability.
Compare Cerebras Models vs Competitors
See detailed cost comparisons between Cerebras models and models from other providers, including side-by-side pricing tables and monthly cost projections.
Frequently Asked Questions: Cerebras API Pricing
What is the cheapest Cerebras model?
The cheapest Cerebras model by input token price is Llama 3.1 8B (Cerebras), priced at $0.10/M for input tokens and $0.10/M for output tokens. It is best suited for high-volume, cost-sensitive workloads where speed and price matter most.
What is the most capable Cerebras model?
Qwen3 235B Instruct (Cerebras) is Cerebras's most capable model, classified as a premium-tier model. It supports a context window of 128,000 tokens and is priced at $0.60/M input / $1.20/M output per million tokens.
How does Cerebras API pricing work?
Cerebras charges per token consumed — separately for input tokens (the text you send) and output tokens (the text the model generates). Prices are listed in USD per 1 million tokens ($/M). Cerebras was founded in 2016 and is headquartered in Sunnyvale, CA. All pricing shown here is sourced from the official Cerebras pricing page.
Does Cerebras offer batch API pricing?
Cerebras's current model lineup does not include batch API pricing at this time. For cost reduction on high-volume workloads, consider prompt caching (where supported) or running multiple concurrent requests with standard pricing.
Do Cerebras models support vision / image inputs?
None of the current Cerebras models in our dataset support vision input. Cerebras models are text-focused. For multimodal workloads requiring image understanding, consider comparing models from other providers using our comparison tool.
Cerebras pricing data last verified: 2026-03-10. Prices may change — verify on the official Cerebras pricing page before production use.