Mistral API review

The Mistral API ranks eighth. Its strongest point is open weights: Mistral Large 4 is described as an open-weight multimodal model and Small 4 is released under Apache 2.0, so teams can start on the hosted API and later run the same model themselves. Mistral's pricing page says most models are priced per million tokens and gives Mistral Large as the example at $0.50 input and $1.50 output, which is low for a flagship. Batch processing cuts prices by 50% and cached input tokens cut input cost by up to 90%. OCR, Voxtral speech and tool APIs are priced per page, minute or call. The docs also list hosted third-party models such as GLM 5.3 and DeepSeek V4.1 Flash. The weaker part is transparency: the public pricing page focuses on the Vibe subscription and gives Mistral Large as its only API price example.

Who it is for

Teams that want European-built models with open weights, plus OCR and speech from the same provider.

Price

Pay per token; Mistral Large at $0.50 per 1M input tokens.

For example, Mistral Large costs $0.5 /M tokens in and $1.5 /M tokens out.
Quoted from the vendor pricing page on November 7, 2026.

Key features

• Mistral Large 4, an open-weight general-purpose multimodal model
• Mistral Small 4, combining instruct, reasoning and coding in one model
• OCR priced per 1,000 pages and Voxtral speech priced per minute
• Batch processing at 50% off and cached input at up to 90% off
• Hosted third-party models such as GLM 5.3 and DeepSeek V4.1 Flash

Strengths and weaknesses

• Strength: Open weights on flagship models allow self-hosting later
• Strength: Mistral Large at $0.50 input and $1.50 output per 1M tokens is low for a flagship
• Strength: Document OCR and speech sit alongside the language models
• Weakness: The public pricing page focuses on the Vibe subscription and gives one API price example
• Weakness: Fewer published speed figures than Groq or Fireworks AI
• Weakness: Smaller tool ecosystem than OpenAI or Google

Key facts

• Flagship price: Mistral Large, $0.50 input and $1.50 output per 1M tokens
• Batch discount: 50%
• Caching: Up to 90% off input for repeated prompts

Where it ranks

Mistral API is No. 8 in our Best LLM APIs verdict.