Mistral API pricing: what each model costs per 1M tokens
Standard, batch and priority prices for Mistral’s current models, cached input rates, how billing works and a worked monthly cost example.
Last updated October 22, 2026
How much does the Mistral API cost?
On October 22, 2026, Mistral Large 4 cost $1.36 per 1M input tokens and $4.18 per 1M output tokens, Mistral Medium 3.5 cost $1.50 and $7.50, Mistral Small 4 cost $0.15 and $0.60, and Ministral 3 3B cost $0.10 both ways. Batch processing halves these prices and cached input costs about a tenth.
What does each Mistral model cost per 1M tokens?
These are the standard prices for Mistral’s flagship and code models from its API pricing page. Mistral’s models overview describes Mistral Medium 3.5 as its frontier-class multimodal model for agentic and coding use, Mistral Large 4 as a general-purpose open-weight multimodal model, and Mistral Small 4 as a hybrid instruct, reasoning and coding model.
| Model | Input | Cached input | Output |
|---|---|---|---|
| Mistral Large 4 | $1.36 | $0.14 | $4.18 |
| Mistral Large 3 | $0.50 | $0.05 | $1.50 |
| Mistral Medium 3.5 | $1.50 | $0.15 | $7.50 |
| Mistral Small 4 | $0.15 | $0.015 | $0.60 |
| Ministral 3 14B | $0.20 | $0.02 | $0.20 |
| Ministral 3 8B | $0.15 | $0.015 | $0.15 |
| Ministral 3 3B | $0.10 | $0.01 | $0.10 |
| Codestral | $0.30 | $0.03 | $0.90 |
Prices checked October 22, 2026. Providers change prices often; confirm on Mistral’s API pricing page.
The same page lists Z.ai GLM 5.3, a third-party model hosted by Mistral, at $1.40 input, $0.14 cached input and $4.40 output, and prices OCR, transcription and speech models per page, minute or character. Note that the example on mistral.ai/pricing quotes “Mistral Large” at $0.5 in and $1.5 out, which matches Mistral Large 3, not Large 4; check which version you call.
How does Mistral API billing work?
- Input and output are billed separately. Mistral prices most models per 1M tokens, counting your prompts as input and the responses as output.
- Cached input is cheaper. Mistral’s pricing FAQ says cached input tokens reduce input cost by up to 90% for repeated prompts; the table shows cached input at about a tenth of the input price.
- Three tiers. Standard, Batch (50% off) and Priority (1.75× standard on the flagship models).
- Some APIs use other units. OCR is priced per 1,000 pages, speech models per minute or per million characters, and tool APIs per call. Voice agents are billed per minute of conversation.
- Usage tracking. Mistral says its Studio dashboard breaks down API usage and costs by service.
What would 1M requests a month cost on Mistral?
Take 1,000,000 requests a month, each with 1,000 input tokens and 300 output tokens: 1,000M input tokens and 300M output tokens, with no caching.
Monthly cost = 1,000 × input price + 300 × output price (prices per 1M tokens)
| Model and tier | Calculation | Monthly cost |
|---|---|---|
| Mistral Medium 3.5, standard | 1,000 × $1.50 + 300 × $7.50 | $3,750 |
| Mistral Large 4, standard | 1,000 × $1.36 + 300 × $4.18 | $2,614 |
| Mistral Small 4, standard | 1,000 × $0.15 + 300 × $0.60 | $330 |
| Mistral Small 4, batch | 1,000 × $0.075 + 300 × $0.30 | $165 |
| Ministral 3 3B, standard | 1,000 × $0.10 + 300 × $0.10 | $130 |
Count your real prompt sizes with the token calculator; token counts vary between providers because each uses its own tokenizer.
How can you cut your Mistral API bill?
| Model | Batch input | Batch cached input | Batch output |
|---|---|---|---|
| Mistral Large 4 | $0.68 | $0.07 | $2.09 |
| Mistral Large 3 | $0.25 | $0.025 | $0.75 |
| Mistral Medium 3.5 | $0.75 | $0.075 | $3.75 |
| Mistral Small 4 | $0.075 | $0.0075 | $0.30 |
| Ministral 3 14B | $0.10 | $0.01 | $0.10 |
| Ministral 3 8B | $0.075 | $0.0075 | $0.075 |
| Ministral 3 3B | $0.05 | $0.005 | $0.05 |
Prices checked October 22, 2026. Providers change prices often; confirm on Mistral’s API pricing page.
- Batch processing. Mistral says batch processing reduces the price by 50% for high-volume work.
- Prompt caching. Repeated prompt content billed as cached input costs about a tenth of normal input.
- Smaller models. Mistral Small 4 costs a tenth of Mistral Medium 3.5 per input token and less than a twelfth per output token. Mistral itself suggests Small for cost-sensitive projects.
- Avoid Priority unless you need it. Priority costs 1.75× standard, for example $2.38 input and $7.315 output on Mistral Large 4.
To estimate what response caching and model routing could save on your traffic, try the LLM gateway savings calculator.
How does Perimattic LLM Gateway help with Mistral costs?
Perimattic LLM Gateway sits between your application and model providers. Its features:
- One OpenAI-compatible API in front of Mistral, OpenAI, Anthropic, Google, DeepSeek and open models.
- Routing rules that send each request to a model by cost, latency or task, with automatic fallback when a provider errors or rate-limits.
- Exact-match and semantic caching of responses.
- Budgets and rate limits per team, API key or customer, with cost tracking and alerts.
Get started: Sign up and our team will walk you through Perimattic LLM Gateway for your workloads.
Where do these Mistral prices come from?
All prices on this page come from Mistral’s API pricing page, with plan and billing details from mistral.ai/pricing and model descriptions from Mistral’s models overview. They are standard prices in US dollars, before tax; Mistral also lists euro prices and regional inference. Prices checked October 22, 2026; providers change prices often — confirm on the provider’s page before you rely on them. Perimattic is not affiliated with Mistral AI.
Compare Mistral with other providers
Mistral API pricing, answered
Which Mistral model should I use?
Mistral’s pricing FAQ suggests Mistral Medium for most tasks and coding, and Mistral Small for cost-sensitive projects; its docs feature Mistral Large 4 as a general-purpose multimodal model. Medium 3.5 costs more per output token than Large 4, so test both on your own prompts rather than assuming the larger name costs more.
Can I self-host Mistral models instead of paying per token?
Mistral says you can self-host its models. Its pricing FAQ notes that open-weight models are Apache 2.0 licensed for research and individual use, while commercial deployments require a Mistral license. Check each model’s licence on Mistral’s models page, and remember that self-hosting swaps token fees for GPU and operations costs.
Does Mistral charge for chat subscriptions and the API separately?
Yes. Mistral’s Free, Pro and Team plans cover its Vibe apps, with some API credits included on paid plans, while the API is billed per token. Most models are priced per 1M input and output tokens; OCR is priced per 1,000 pages and speech models per minute or per character.
Want one API for every model, with costs under control?
Sign up to tell us which models and providers you use, and our team will get in touch to set up the LLM Gateway for your stack.
Sign up