Skip to main content
Meo

Mistral: Ministral 3 14B 2512

Mistral: Ministral 3 14B 2512 is mistralai's model. On the un-leaked meo benchmark it ranks #53 of 82 with 56.5% accuracy and an Effective Value (𝕍) of 0.0036 β€” earning a correct answer for about $0.0007 in 25s each.

as of 2026-10-18

meo rank
#53
of 82
Accuracy
56.5%
overall
𝕍 index (top=100)
0.086
raw 𝕍 0.0036 Β· N=10
$ / correct
$0.0007
25s each

Per-domain accuracy

139 correct of 250 questions asked across 11 domains Β· answer rate 100% Β· macro accuracy 56.9%. Headline accuracy is micro-averaged per meo’s methodology; per-domain question counts (correct/asked) are shown so you can judge significance.

Tape machine
87%13/15
Base-rate bias
83%34/41
Logic / math / CS
71%24/34
Framework bias
67%2/3
Theory of mind
67%2/3
CSP (unsatisfiable)
56%9/16
Perceptual illusions
45%46/102
Regex automaton
44%7/16
Long arithmetic
13%2/15
Watson-Glaser
0%0/4
State tracking
0%0/1

Efficiency

Cost per correct
$0.0007
Seconds per correct
25
Tokens per correct
2,286
Correct answers
139 / 250
Total benchmark cost
$0.10
Avg tokens / item
1,271

Model details

Released
2025-12-02
Weights
Open
Context
262K
Max output
210K
Knowledge cutoff
β€”
Price in
$0.20/M
Price out
$0.20/M
Modality
text+image->text
View on OpenRouter β†—Hugging Face β†—

Sources: openrouter Β· meo

FAQ

How accurate is Mistral: Ministral 3 14B 2512?

Mistral: Ministral 3 14B 2512 scores 56.5% overall on the meo benchmark (rank #53 of 82), an un-leaked private holdout graded objective-first.

What is Mistral: Ministral 3 14B 2512's Effective Value?

Its Effective Value (𝕍) is 0.0036 at N=10, Ο‰=1, Ξ΄=1.5 β€” a fused metric of accuracy, speed, cost, and error-cascade. 𝕍 rankings shift with task depth.

How much does Mistral: Ministral 3 14B 2512 cost per correct answer?

About $0.0007 per correct answer (2,286 tokens, 25 seconds each) at benchmark-time pricing via OpenRouter.

Is Mistral: Ministral 3 14B 2512 open weights?

Yes β€” Mistral: Ministral 3 14B 2512 is released with open weights.

Benchmark updates affecting this model

  • 2026-06-09majorlogic_math_cs ground truth was broken β€” culled and rebuilt with solver-verified items

    A 5-lab frontier jury revealed that 29 of 37 LLM-authored logic_math_cs items had broken ground truth: on most, five independent frontier models unanimously concluded the clue sets were contradictory while the stored answer claimed a number, and on eight, all five agreed on the SAME different number (author arithmetic errors). Models were being marked wrong for being right. The 29 were retired (full audit published in the repo) and the domain rebuilt with generator-as-oracle items: an embedded constraint solver PROVES each puzzle has exactly one solution (or provably none), verified by an independent exhaustive enumerator. A 3-round difficulty study found frontier models saturate consistent human-scale logic puzzles, so items use anonymous global constraints and carry a measured difficulty gradient.

  • 2026-06-08majorCorrected base_bias grading (LLM-equivalence judge fix)

    The LLM-equivalence grader uses a reasoning model whose verdict token budget was too small, so its YES/NO answer was emitted as empty and silently counted as a failure. This understated the base_bias domain to ~7% for every model when the true accuracy is ~85-90%. We raised the judge budget, excluded its reasoning tokens from the verdict, switched to verbosity-robust conclusion matching, and re-graded the whole board.

Full benchmark changelog β†’

Embed this ranking

Show Mistral: Ministral 3 14B 2512’s meo rank on your site β€” the badge links back here.

Related: all Mistral AI models Β· Tape-machine simulation domain Β· best Effective Value Β· biggest movers Β· methodology Β· full leaderboard.