Mistral AI · llm model

Ministral 3 14B

Release Ruler for Ministral 3 14B: released Dec 2, 2025, $0.2 per 1M input tokens, open weights. Each thin stem on the ruler is another model we track.

Ministral 3 14B is a Mistral AI language model released Dec 2, 2025; Mistral AI's newest language model in our dataset is Mistral Large 4. Open weights (Apache 2.0). Mistral's largest Ministral 3 model, released Dec 2, 2025 under Apache 2.0 and optimized for local deployment. Verified against vendor sources on Oct 26, 2026.

Reported Last verified · All models · JSON · Markdown

  • ReleasedDec 2, 2025
  • Context window256K256,000 tokens
  • LicenceOpen weightsApache 2.0

Modalities not verified

Key facts

Ministral 3 14B: key facts with sources
FactValueSource
VendorMistral AIn/a
ReleasedDec 2, 2025source
Statusgenerally availablen/a
API idministral-14b-2512source
Context window256,000 tokens (256K)source
LicenceApache 2.0 (open weights)source
Pricing$0.2 in / $0.2 out per 1M tokenssource

Where Ministral 3 14B sits in our data

  • Released Dec 2, 2025 (week 49 of 2025), in December 2025. It is the only model in our data released that month.
  • Of the 60 models in our data with a published context window, Ministral 3 14B has the number 51 largest (256K tokens). Our data is not every model on the market.
  • Mistral AI tools we cover: Mistral Le Chat.
  • All releases by date: AI model releases by month · new AI models this week. Choosing for code: best LLM for coding.

How Ministral 3 14B differs from Mistral Medium 3.5

Read from the recorded facts of the two models, not from a head-to-head test. We have not run either.

  • Listed price is $0.2 in and $0.2 out, against $1.5 in and $7.5 out for Mistral Medium 3.5 (per 1M tokens).
  • Both list a 256K-token context window.
  • Ministral 3 14B came out 147 days before Mistral Medium 3.5.
  • Facts for the other model: Mistral Medium 3.5.

What we tested

Nothing yet. We have not run Ministral 3 14B ourselves, so every fact on this page is Reported from the sources below. When we do test it, the scope will be stated here and in our test logs. See methodology.

Sources

More from Mistral AI

All 88 models