MistralReleased on Dec 4, 2025

MiniStral 3 (14B Instruct 2512): Benchmarks, Pricing & Context Window

MiniStral 3 (14B Instruct 2512) is a language model from Mistral, released in December 2025, with multimodal input.

A balanced model in the Ministral 3 family, Ministral 3 14B is a powerful, efficient tiny language model with vision capabilities. This model is the instruct post-trained version in FP8, fine-tuned for instruction tasks, making it ideal

MiniStral 3 (14B Instruct 2512) API

API access coming soon

MiniStral 3 (14B Instruct 2512) will be available through our gateway shortly.

MiniStral 3 (14B Instruct 2512) examples

Recent arena outputs from MiniStral 3 (14B Instruct 2512), picked from the highest-ranked matchups.

MiniStral 3 (14B Instruct 2512) license

MiniStral 3 (14B Instruct 2512) is released under the Apache 2.0 license, which permits commercial use, has 14.0B parameters.

License
Apache 2.0
Commercial use allowed
Parameters
14.0B

Apache License 2.0 - allows commercial use

FAQ

Common questions about MiniStral 3 (14B Instruct 2512).

When was MiniStral 3 (14B Instruct 2512) released?

MiniStral 3 (14B Instruct 2512) was released on December 4, 2025 by Mistral. This is the official MiniStral 3 (14B Instruct 2512) release date tracked on LLM Stats.

Is MiniStral 3 (14B Instruct 2512) available via API?

Yes, MiniStral 3 (14B Instruct 2512) is available via API. See the official documentation for authentication and endpoint details.

How many parameters does MiniStral 3 (14B Instruct 2512) have?

MiniStral 3 (14B Instruct 2512) has 14 billion parameters. It ships as an open-weight model, so you can download and run it on your own hardware.

Who created MiniStral 3 (14B Instruct 2512)?

MiniStral 3 (14B Instruct 2512) was created by Mistral.

What is the license for MiniStral 3 (14B Instruct 2512)?

MiniStral 3 (14B Instruct 2512) is released under the Apache 2.0 license. This is an open-source / open-weight license that permits self-hosting.

Is MiniStral 3 (14B Instruct 2512) multimodal?

Yes, MiniStral 3 (14B Instruct 2512) is multimodal and can accept both text and images as input.

Where is the MiniStral 3 (14B Instruct 2512) paper or technical report?

MiniStral 3 (14B Instruct 2512) has a paper or technical report available at https://mistral.ai/news/mistral-3. Use that source for architecture, training, release and evaluation details.