Our first flaship models handling instruction-following, reasoning, and coding in a single set of opened-weights.
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
View all PapersA state-of-the-art model, open-weight, with a granular Mixture-of-Experts architecture that fuses instruct, reasoning and agentic skills.
A state-of-the-art, open-weight, general-purpose multimodal model with a granular Mixture-of-Experts architecture.
Different formats and Quantized versions of our Ministral 3 family; 14B/8B/3B Instruct/Reasoning GGUF, 3B Instruct ONNX and 14B/8B/3B Instruct BF16.
Mistral AI Audio models.
-
mistralai/Voxtral-4B-TTS-2603
Text-to-Speech • Updated • 74.5k • 868 -
mistralai/Voxtral-Mini-4B-Realtime-2602
Automatic Speech Recognition • 4B • Updated • 2M • 900 -
mistralai/Voxtral-Small-24B-2507
Audio-Text-to-Text • 24B • Updated • 109k • 505 -
mistralai/Voxtral-Mini-3B-2507
5B • Updated • 304k • 659
A couple of agentic LLMs for software engineering tasks, excelling at using tools to explore codebases, edit multiple files, and power SWE Agents.
A collection of edge models, with Base, Instruct and Reasoning variants, in 3 different sizes: 3B, 8B and 14B. All with vision capabilities.
Synthesized speech evals generated by MistralAI from popular text evaluation datasets to evaluate spoken-language reasoning capabilities of Audio LLMs
Our first flaship models handling instruction-following, reasoning, and coding in a single set of opened-weights.
Mistral AI Audio models.
-
mistralai/Voxtral-4B-TTS-2603
Text-to-Speech • Updated • 74.5k • 868 -
mistralai/Voxtral-Mini-4B-Realtime-2602
Automatic Speech Recognition • 4B • Updated • 2M • 900 -
mistralai/Voxtral-Small-24B-2507
Audio-Text-to-Text • 24B • Updated • 109k • 505 -
mistralai/Voxtral-Mini-3B-2507
5B • Updated • 304k • 659
A state-of-the-art model, open-weight, with a granular Mixture-of-Experts architecture that fuses instruct, reasoning and agentic skills.
A couple of agentic LLMs for software engineering tasks, excelling at using tools to explore codebases, edit multiple files, and power SWE Agents.
A state-of-the-art, open-weight, general-purpose multimodal model with a granular Mixture-of-Experts architecture.
A collection of edge models, with Base, Instruct and Reasoning variants, in 3 different sizes: 3B, 8B and 14B. All with vision capabilities.
Different formats and Quantized versions of our Ministral 3 family; 14B/8B/3B Instruct/Reasoning GGUF, 3B Instruct ONNX and 14B/8B/3B Instruct BF16.
Synthesized speech evals generated by MistralAI from popular text evaluation datasets to evaluate spoken-language reasoning capabilities of Audio LLMs