Skip to contentT/TTokens or Towers

Compare

LeaderboardEvery tracked model, ranked and priced side by sideBenchmarksWhat each test measures and who leads itPricesAPI, media, and hardware rates with a cost calculator

Plan

Stack plannerSix questions to a priced local and hosted stackShortlistOur picks for subscriptions, APIs, and hardwareEnterpriseSeat plans against self-hosting for a whole team

Build

HardwareGPUs and workstations, and which models they fitHostsWhere to rent open-weight inference, and for how muchHarnessesCoding agents and the models they drive

Lab

FindingsResults from our own evaluationsScanAudit a repo’s AI usage in the browserMethodWhere the numbers come from and how we check them
Data checked September 25, 2026
Data checked September 25, 2026
Plan a stack
T/TTokens or Towers

An independent guide to AI models, what they cost, and the hardware to run them. No account, no API key, no sponsors.

Compare

  • Leaderboard
  • Benchmarks
  • Prices

Plan

  • Stack planner
  • Shortlist
  • Enterprise

Build

  • Hardware
  • Hosts
  • Harnesses

Lab

  • Findings
  • Scan
  • Method
Data checked September 25, 2026 · Every number links to its sourceRSS · Source on GitHub
Leaderboard/Mistral
Mistral

Mistral · Ministral 3

Ministral 3 3B

Open weightsGenerally availableApache-2.0

Smallest Ministral 3 model for on-device use.

AA Intelligence
4.8#141 of 142
Input / output per 1M
$0.10/ $0.10
Output speed
209tok/s
Context window
256K tokens

Composite indexes

  • Artificial Analysis Intelligence Index#141 of 142
    5independentIntelligence Index v4.3; non-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Opus 5.5 at 58

Coding

  • SciCode#93 of 94
    15.3%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Opus 5.5 at 66.9%

  • Terminal-Bench 4.0#80 of 88
    0%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Mythos 5.1 at 60.9%

  • Terminal-Bench 2.x#103 of 103
    0%independentTerminal-Bench 2.1; non-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Fable 5.1 at 91.4%

  • LiveCodeBench#47 of 56
    24.7%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Fugu at 92.9%

Agents and tool use

  • τ²-Bench#60 of 88
    24.9%independenttelecom; non-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: GLM-5.2 at 99.1%

  • GDPval-AA#96 of 103
    15independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Opus 5.5 at 1846

Reasoning and knowledge

  • GPQA Diamond#135 of 142
    35.8%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: GPT-6 Astra at 96.1%

  • Humanity’s Last Exam#119 of 147
    5.4%independentno tools; non-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Opus 5.5 at 61.4%

  • CritPt#108 of 139
    0%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: GPT-5.6 Sol at 32.3%

Vision and long context

  • AA Long Context Reasoning#119 of 139
    17%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Kimi K3 at 88.7%

  • MMMU-Pro#69 of 72
    38.1%independentnon-reasoningArtificial Analysis: Ministral 3 3B ↗

    Leader: Claude Opus 5.5 at 87.7%

Other published results

ResultScoreReported bySource
aime-2025 · AIME 2025; non-reasoning22%independentArtificial Analysis: Ministral 3 3B ↗
Terminal-Bench Hard (AA) · non-reasoning0%independentArtificial Analysis: Ministral 3 3B ↗
IFBench · non-reasoning26.8%independentArtificial Analysis: Ministral 3 3B ↗
tau2-Bench Banking (AA) · non-reasoning4.7%independentArtificial Analysis: Ministral 3 3B ↗
AA-Omniscience Index · non-reasoning-64 indexindependentArtificial Analysis: Ministral 3 3B ↗

Specs

Released
December 2, 2025
Parameters
3B
Size class
Under 10B
Input
text, image
Output
text
Context
256K tokens
Max output
Not published
Blended 70/30
$0.10 / 1M
First token
0.70s
  • Price: Mistral docs model page ↗ Mistral API list price per 1M tokens.
  • Speed: Artificial Analysis leaderboard (Ministral 3 3B, median across providers or first-party as shown) ↗

Will it fit your budget?

The planner prices this model against local hardware and other APIs for your workload.

Plan a stack →

Compare with

  • CohereTiny Aya Global4.8
  • Ai2Molmo 2 8B5
  • Swiss AI InitiativeApertus 70B Instruct5.1
  • AI21 LabsJamba 1.7 Mini5.2