Mistral AI review

Mistral AI is the European option on this list and the one most focused on open weights. Its pricing FAQ gives Mistral Large at $0.50 per 1M input tokens and $1.50 per 1M output tokens, a low price for a flagship model, and says batch processing cuts prices by 50% and cached input by up to 90%. The models overview lists Mistral Large 4 as an open-weight general-purpose multimodal model and Mistral Small 4 under Apache 2.0, so teams can move a workload to their own hardware later. The same platform carries OCR, Voxtral speech and moderation models, and Mistral Studio tracks API calls and spend. Mistral also sells consumer and coding plans under the Vibe brand. The weak spots for a developer comparing APIs are documentation gaps: the pages we checked did not state context windows for the main models or publish rate limit tiers, and several model versions have retirement dates within months.

Who it is for

Teams that want a European vendor with open-weight models they can also self-host.

Price

Pay as you go; Mistral Large costs $0.50 per 1M input tokens and $1.50 per 1M output 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 under Apache 2.0 for cost-sensitive work
• Batch processing at 50% off
• Cached input tokens up to 90% cheaper
• OCR, Voxtral speech and moderation models on the same platform

Strengths and weaknesses

• Strength: A flagship model at $0.50 in and $1.50 out per 1M tokens
• Strength: Open-weight licenses allow self-hosting
• Strength: Usage tracking in the Mistral Studio dashboard
• Weakness: Context windows are not stated on the models overview we checked
• Weakness: Published rate limit tiers were not available when we checked
• Weakness: Several model versions retire within months

Key facts

• Mistral Large: $0.50 in, $1.50 out per 1M tokens
• Batch discount: 50%
• Cached input: Up to 90% cheaper
• Licenses: Apache 2.0 (Small 4), Mistral open weight (Large 4)

Where it ranks

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