Mistral AI Models

Approved models

47

Records published from this collection

Provider information

lab

Review approach

Verify each model’s source, terms, capabilities, and current availability

Mistral publishes hosted models and downloadable checkpoints for text, vision, documents and audio. These products have different interfaces and licences, so choose a model for the actual input and output your application needs.

General assistants and coding

Mistral Large 4 is a public-preview multimodal model with a one-million-token context allowance. Its promotional API pricing is a dated offer. Medium 3.5 and Small 4 provide useful alternative deployment baselines. Compare a repository fix or document task before routing every request to the largest model.

Purpose-built interfaces

OCR 4.1 handles document extraction; evaluate layout and field accuracy. Shieldstral 1 is a moderation classifier whose policy and error rates need a separate evaluation. Voxtral Mini Transcribe 2 turns audio into text, while Voxtral TTS produces speech. These are different endpoints, with billing units that should not be forced into a text-token comparison.

Check the deployment terms

Mistral’s downloadable releases do not all share one licence. Keep the checkpoint licence, API model ID, region and processing tier with your decision. For each candidate, track the result a user can accept, retries and review time. A published benchmark can help shortlist a model, but it cannot establish the result for your integration.

Publisher resources

Current publisher documentation

Mistral AI model list

Review each record individually; collection membership does not imply a recommendation or guarantee of availability.

47 models
Mistral Large 4
Mistral AI
66.61M$0.68$2.09—
Shieldstral 1.0
Mistral AI
19.732K———
Mistral OCR 4.1
Mistral AI
14.5————
Mistral Medium (latest)
Mistral AI
69.6262.14K$1.50$7.50—
Mistral Medium 3.5
Mistral AI
69.6256K$1.50$7.50—
Voxtral TTS
Mistral AI
13.9————
Mistral Small (latest)
Mistral AI
64.3256K$0.15$0.60—
Mistral Small 4
Mistral AI
64.3256K$0.15$0.60—
Devstral 2
Mistral AI
43.9262.14K$0.40$2.00—
Devstral Small 2
Mistral AI
—262.14K———
Devstral 2 (latest)
Mistral AI
43.8262.14K$0.40$2.00—
Ministral 14B
Mistral AI
—262.14K———
Ministral 3 14B
Mistral AI
—262.14K———
Ministral 3 3B
Mistral AI
—262.14K———
Ministral 3 8B
Mistral AI
—262.14K———
Mistral Large 3
Mistral AI
46.7262.14K$0.50$1.50—
Magistral Small 1.2
Mistral AI
—131.07K———
Mistral Medium 3.1
Mistral AI
—131.07K———
Codestral 25.08
Mistral AI
—128K———
Voxtral Mini 3B 2507
Mistral AI
—32.77K———
Voxtral Small (latest)
Mistral AI
—32K———
Devstral Medium
Mistral AI
38.6128K$0.40$2.00—
Devstral Small
Mistral AI
38.6128K$0.10$0.30—
Mistral Small 3.2
Mistral AI
39.0128K$0.10$0.30—
Magistral Small
Mistral AI
—131.07K———
Mistral Medium 3
Mistral AI
43.4131.07K$0.40$2.00—
Pixtral Large (25.02)
Mistral AI
—128K———
Magistral Medium (latest)
Mistral AI
48.5128K$2.00$5.00—
Mistral Small 3.1 24B
Mistral AI
—128K———
Saba
Mistral AI
—32.77K———
Mistral Small 3
Mistral AI
—32.77K———
Mistral Large 2407
Mistral AI
—131.07K———
Mistral Large 2.1
Mistral AI
33.0131.07K$2.00$6.00—
Mistral Large (latest)
Mistral AI
46.7262.14K$0.50$1.50—
Pixtral Large (latest)
Mistral AI
42.2128K$2.00$6.00—
Ministral 3B
Mistral AI
—128K———
Ministral 8B Instruct
Mistral AI
—131.07K———
Pixtral 12B
Mistral AI
41.9128K$0.15$0.15—
Mistral Nemo
Mistral AI
36.6128K$0.15$0.15—
Codestral (latest)
Mistral AI
32.3256K$0.30$0.90—
Codestral-22B-v0.1
Mistral AI
—32.77K———
Mixtral 8x22B Instruct
Mistral AI
—65.54K———
Mistral Large
Mistral AI
—128K———
Mistral Embed
Mistral AI
—8K———
Leanstral 1.5 119B A6B
Mistral AI
40.2262.14K———
Leanstral 2603
Mistral AI
40.2262.14K———
Voxtral Mini Transcribe 2
Mistral AI
5.0————

1–47 of 47 models · select a column to sort

About this order: Recommended shows our editorial shortlist first, followed by recent releases. The spec score is a 0–100 figure SyncDev calculates from each vendor's published specifications — context window, reasoning support, input modalities, tool calling, maximum output and how recently the model shipped. It measures breadth of capability, not benchmark performance, so a higher-scoring model is not automatically the better choice for your task.

About Mistral AI models

It groups source-backed directory records for research and comparison. Review each model page and its primary documentation because collection membership alone does not prove quality, availability, or suitability.