Skip to main content

    Mistral AI

    Ministral 3

    Mistral's Dec 2, 2025 compact open-weight family — 3B, 8B, and 14B sizes, each in Base/Instruct/Reasoning variants, all with vision. 256K context, Apache 2.0; built for edge deployment and fine-tuning.

    256K tokens
    Multimodal
    Pro
    Provider
    Mistral AI
    Context
    256K
    Catalog · As of Aug 11, 2026Stale · 31d old
    yno subscription tier
    Pro
    Released
    Dec 2, 2025
    Speed
    Fast
    Reasoning
    Basic
    Modality
    Multimodal

    Benchmarks

    Quality

    AA Intelligence Index
    6points
    Artificial Analysis · Date unknownAge unknown
    AA Coding Index
    14.4points
    Artificial Analysis · Date unknownAge unknown
    LiveCodeBench
    35.1%
    Artificial Analysis · Date unknownAge unknown
    GPQA Diamond
    57.2%
    Artificial Analysis · Date unknownAge unknown

    Speed and latency

    Output speed
    85.97tokens/s
    Configuration: Prompt length: 1,000 · Parallel queries: 1
    Artificial Analysis · Date unknownAge unknown
    Time to first token
    0.58s
    Configuration: Prompt length: 1,000 · Parallel queries: 1
    Artificial Analysis · Date unknownAge unknown
    Time to first answer token
    0.58s
    Configuration: Prompt length: 1,000 · Parallel queries: 1
    Artificial Analysis · Date unknownAge unknown

    API pricing

    API input cost
    $0.20/ 1M tokens
    Artificial Analysis · Date unknownAge unknown
    API output cost
    $0.20/ 1M tokens
    Artificial Analysis · Date unknownAge unknown

    Benchmark variant: Ministral 3 14B

    AA data retrieved Sep 11, 2026 · Artificial Analysis

    Indexes use points; evaluations use accuracy percentages. API measurements do not measure yno application performance. Methodology

    Capabilities

    • Vision
    • Open Source
    • Edge
    • Fine-tuning

    Use cases

    • Edge deployment
    • Cost-sensitive apps
    • Custom fine-tuning
    • Vision tasks

    Strengths

    • 3B/8B/14B family
    • Apache 2.0
    • Vision on every size
    • 256K context

    Best for

    Developers needing a small, capable, customizable vision model

    Use Ministral 3 in yno.ai

    No credit card required

    Related models

    Frequently asked

    What is Ministral 3?
    Mistral's Dec 2, 2025 compact open-weight family — 3B, 8B, and 14B sizes, each in Base/Instruct/Reasoning variants, all with vision. 256K context, Apache 2.0; built for edge deployment and fine-tuning.
    How much does Ministral 3 cost in yno.ai?
    Ministral 3 is available on the Pro tier of yno.ai.
    What can Ministral 3 do?
    Ministral 3 is best for Developers needing a small, capable, customizable vision model. Its main capabilities include Vision, Open Source, Edge, Fine-tuning.
    How does Ministral 3 compare to other models?
    Ministral 3 excels at 3B/8B/14B family and is recommended for Edge deployment, Cost-sensitive apps, Custom fine-tuning, Vision tasks. See related models below.
    How do I use Ministral 3 in yno.ai?
    Sign up for yno.ai, select Ministral 3 from the model picker, and start chatting.