Inference
EN
Mistral

Mistral Small 3.1

Mistral Small 3.1 is a standard model from Mistral. 20 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Standard
  • Input: text
  • Released 2025-03
RankedEnough results for an index.
IndexAll categories, out of 100.24.2±4.1Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.28/sMedian 67/s
First tokenSeconds to the first output token.1.0 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.128KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

24.01B parameters
16-bit
51.4 GB
8-bit
27.8 GB
4-bit
16.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 24 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted26.7208 / 218
CodingCode writing and repair.3 counted25.6202 / 209
ReasoningLogic problems and puzzles.2 counted29.7209 / 227
MultimodalTasks with images and text.2 counted28.882 / 84
KnowledgeFacts and expert knowledge.6 counted25.7255 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted17.0—
MathMath problems.4 counted16.7162 / 168

Results

20 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Evaluating Large Language Models Trained on CodeCoding88.4%8 / 22
MATH level 5Math46.8% ±1.251 / 88
Critical Physics TasksReasoning0.0%150 / 214
Massive Multitask Language UnderstandingKnowledge80.6%24 / 34
GPQA diamond · Epoch AIKnowledge44.7% ±2.3164 / 217
GDPval-AA normalizedAgentic0.0%101 / 133
Massive Multi-discipline Multimodal UnderstandingMultimodal64.0%22 / 28
τ²-Bench Tool-Agent-User EvaluationAgentic25.1%120 / 149
Massive Multitask Language Understanding ProfessionalKnowledge66.8%61 / 75
OTIS Mock AIME 2024-2025Math4.9% ±1.5170 / 200
Artificial Analysis Humanity's Last ExamKnowledge4.3%190 / 222
Measuring Short-Form Factuality in Large Language ModelsKnowledge10.4%12 / 14
Artificial Analysis Long Context ReasoningReasoning22.3%191 / 217
Graduate-Level Google-Proof Q&AKnowledge46.0%127 / 141
MMMU ProMultimodal60.1% ±1.283 / 92
Artificial Analysis GPQA DiamondKnowledge45.4%182 / 199
MATH 500Math68.4% ±2.154 / 59
Artificial Analysis SciCodeCoding27.8%105 / 114
Artificial Analysis IFBenchInstruction29.9%135 / 146
Artificial Analysis Omniscience AccuracyKnowledge15.1%152 / 164
MMLU ProKnowledge66.0% ±0.5126 / 135
GPQA Diamond · Vals AIKnowledge44.2% ±2.5127 / 135
Massive Multi-discipline Multimodal Understanding ProMultimodal49.3%50 / 53
LiveCodeBenchCoding31.8% ±1.1137 / 142
AIMEMath3.5% ±0.292 / 95
MortgageTaxFinance61.2% ±1.062 / 97
LegalBenchLegal69.2% ±0.5127 / 149
MedQAHealthcare69.1% ±0.482 / 95
CorpFin v2Finance44.2% ±1.0118 / 131
TaxEval v2Finance58.3% ±1.0133 / 143
MGSMMultilingual84.2% ±2.269 / 74

Sources

BenchLMCC BY-NC 4.0Vals AINo licence statedEpoch AICC BY

History

5 releases
Index25.0 → 24.2
23 Oct 202612 Nov 2026

Questions

How good is Mistral Small 3.1?
Its index is 24.2 out of 100. That is place 271 of 283 ranked models.
What is Mistral Small 3.1 best at?
Reasoning: 29.7 out of 100, place 209 of 227.
How fast is Mistral Small 3.1?
28 output tokens per second, 1.0 s to the first token.
What is the context window of Mistral Small 3.1?
128K tokens.
Is Mistral Small 3.1 open source?
Yes. Its weights are published.
When was Mistral Small 3.1 released?
March 2025.

Badge

Mistral Small 3.1 on Inference360: place 271 of 283, index 24.2
HTML
<a href="https://inference360.ai/models/mistral-small-3-1"><img src="https://inference360.ai/badges/mistral-small-3-1.svg" alt="Mistral Small 3.1 on Inference360: place 271 of 283, index 24.2" height="20"></a>
Markdown
[![Mistral Small 3.1 on Inference360: place 271 of 283, index 24.2](https://inference360.ai/badges/mistral-small-3-1.svg)](https://inference360.ai/models/mistral-small-3-1)

Compare with

Claude 3.5 Haiku24.1Command A23.9Llama 3.1 70B24.8Mistral Medium 327.1Mistral Large 227.4

More from Mistral

All models
Mistral Small 3.1 24BUnrankedSabaUnrankedMistral Medium 327.1Devstral Small28.2Magistral Medium33.7