Side-by-side comparison

Compare AI Models

Pick two to five models. Every cell comes from a verified record; anything unverified says so. There is no overall winner — the right choice depends on your task, budget and technical needs.

CriterionMistral Medium 3.5Mistral Large 4Mistral Small 4
ProviderMistral AIMistral AIMistral AI
Release dateApr 28, 2026Oct 6, 2026Mar 16, 2026
StatusGenerally available Preview PreviewGenerally available
PurposeMultimodalFlagshipFast / lightweight
Inputs → outputstext, image → texttext, image → texttext → text
Context windowNot documented1,000,000 tokens256,000 tokens
API input price$1.50 per 1M tokens$0.68 per 1M tokens$0.15 per 1M tokens
API output price$7.50 per 1M tokens$2.09 per 1M tokens$0.60 per 1M tokens
Example: 1,000,000 in + 200,000 out tokens$3$1.10$0.27
Licenceopen-weight (Modified MIT)open-weight (licence not yet published)open-source (Apache 2.0)
Last verifiedOct 19, 2026Oct 19, 2026Oct 19, 2026

The example cost multiplies the published per-1M-token list prices by the token counts above. It ignores caching, batch discounts, long-context surcharges and tax, and is shown only when both rates are verified.

How to choose

Start from the rows that matter for your job: a free plan and learning curve for casual use; API access, rates and data terms for building products; context window and modalities for long documents or images. Provider claims are labelled as such; we do not rank.

Ready-made comparisons