Mistral's flagship open-weight models — built for breadth. Frontier-grade reasoning, multimodal input, code generation, and agentic capabilities.
AI & ML interests
None defined yet.
Recent Activity
View all activity
Frontier model handling instruction-following, reasoning, and coding in a single set of opened-weights.
A compact 3B-parameter, policy-adaptive multimodal safety classifier.
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 • 2.63k • 956 -
mistralai/Voxtral-Mini-4B-Realtime-2602
Automatic Speech Recognition • 4B • Updated • 1.54M • 1.06k -
mistralai/Voxtral-Mini-4B-Realtime-Arabic
Automatic Speech Recognition • 4B • Updated • 2.63k • 55 -
mistralai/Voxtral-Small-24B-2507
Audio-Text-to-Text • 24B • Updated • 6.59k • 556
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.
A couple of agentic LLMs for software engineering tasks, excelling at using tools to explore codebases, edit multiple files, and power SWE Agents.
Synthesized speech evals generated by MistralAI from popular text evaluation datasets to evaluate spoken-language reasoning capabilities of Audio LLMs
Domain-specialist models trained on Forge, Mistral's platform for building custom AI — from pre-training to post-training and reinforcement learning.
Time series foundation models for zero-shot probabilistic forecasting.
-
mistralai/Timestral-209M
Time Series Forecasting • 0.2B • Updated • 28k • 490 -
mistralai/Timestral-123M
Time Series Forecasting • 0.1B • Updated • 9.55k • 47 -
mistralai/Timestral-28M
Time Series Forecasting • 27.8M • Updated • 17.6k • 53 -
mistralai/Timestral-5M
Time Series Forecasting • 5.21M • Updated • 13.4k • 84
An open-source code agent model designed for Lean 4.
A state-of-the-art model, open-weight, with a granular Mixture-of-Experts architecture that fuses instruct, reasoning and agentic skills.
A collection of edge models, with Base, Instruct and Reasoning variants, in 3 different sizes: 3B, 8B and 14B. All with vision capabilities.
Models co-developed in partnership with Morocco's Ministère de la Transition Numérique et de la Réforme Administrative (MTNRA)
A state-of-the-art, open-weight, general-purpose multimodal model with a granular Mixture-of-Experts architecture.
Mistral's flagship open-weight models — built for breadth. Frontier-grade reasoning, multimodal input, code generation, and agentic capabilities.
Domain-specialist models trained on Forge, Mistral's platform for building custom AI — from pre-training to post-training and reinforcement learning.
Frontier model handling instruction-following, reasoning, and coding in a single set of opened-weights.
Time series foundation models for zero-shot probabilistic forecasting.
-
mistralai/Timestral-209M
Time Series Forecasting • 0.2B • Updated • 28k • 490 -
mistralai/Timestral-123M
Time Series Forecasting • 0.1B • Updated • 9.55k • 47 -
mistralai/Timestral-28M
Time Series Forecasting • 27.8M • Updated • 17.6k • 53 -
mistralai/Timestral-5M
Time Series Forecasting • 5.21M • Updated • 13.4k • 84
A compact 3B-parameter, policy-adaptive multimodal safety classifier.
An open-source code agent model designed for Lean 4.
Our first flaship models handling instruction-following, reasoning, and coding in a single set of opened-weights.
A state-of-the-art model, open-weight, with a granular Mixture-of-Experts architecture that fuses instruct, reasoning and agentic skills.
Mistral AI Audio models.
-
mistralai/Voxtral-4B-TTS-2603
Text-to-Speech • Updated • 2.63k • 956 -
mistralai/Voxtral-Mini-4B-Realtime-2602
Automatic Speech Recognition • 4B • Updated • 1.54M • 1.06k -
mistralai/Voxtral-Mini-4B-Realtime-Arabic
Automatic Speech Recognition • 4B • Updated • 2.63k • 55 -
mistralai/Voxtral-Small-24B-2507
Audio-Text-to-Text • 24B • Updated • 6.59k • 556
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.
Models co-developed in partnership with Morocco's Ministère de la Transition Numérique et de la Réforme Administrative (MTNRA)
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.
Synthesized speech evals generated by MistralAI from popular text evaluation datasets to evaluate spoken-language reasoning capabilities of Audio LLMs