Passer au contenu principal

Mistral Small 3.1

MistralMistralOpen WeightApache 2.0 · Usage Commercial

Description

Mistral Small 3 is a 24B-parameter LLM licensed under Apache-2.0. It focuses on low-latency, high-efficiency instruction following, maintaining performance comparable to larger models. It provides quick, accurate responses for conversational agents, function calling, and domain-specific fine-tuning. Suitable for local inference when quantized, it rivals models 2–3× its size while using significantly fewer compute resources.

Date de sortie
2025-03-17
Paramètres
24.0B
Longueur du contexte
33K
Modalités
image, text

Radar de capacités

29
general
25
coding
19
reasoning
31
science
21
agents
60
multimodal

Classements

Domaine#RangScoreSource
Classement codage376
26.0
AA
Classement général390
33.0
AA
Science428
31.0
AA

Scores de benchmarks (LLM Stats)

(LLM Stats (zeroeval))

Biology

GPQANYU + Cohere + Anthropic (2023)46.0%Aut.

Code

HumanEvalOpenAI (2021)88.4%Aut.

Communication

MT-Bench0.83 / 100Aut.
Wild Bench52.2%Aut.

Creativity

Arena Hard87.6%Aut.

Factuality

SimpleQA10.4%Aut.

Finance

MMLU80.6%Aut.
MMLU-Pro66.8%Aut.

General

IFEvalGoogle Research (2023)82.9%Aut.
TriviaQA80.5%Aut.
MBPP0.75 / 100Aut.
MMMU59.3%Aut.

Math

MATH69.3%Aut.

Indices d'évaluation AA

(Artificial Analysis)
Coding Index(Artificial Analysis)
26.3
Intelligence Index(Artificial Analysis)
14.9
Math Index(Artificial Analysis)
3.7
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))
0.7
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))
0.7
Gpqa(NYU + Cohere + Anthropic (2023))
0.5
Ifbench(Google Research (2023))
0.3
Scicode(UIUC + Argonne National Lab (2024))
0.3
Terminalbench V2 1
0.3
Tau2(Sierra + U Toronto + Vector Institute (2025))
0.3
Lcr(Artificial Analysis)
0.2
Livecodebench(UC Berkeley + MIT + Cornell (2024))
0.2
Aime(MAA (Mathematical Association of America))
0.1
Terminalbench Hard(Stanford × Laude Institute (2026))
0.1
Tau Banking
0.1
Hle(Center for AI Safety + Scale AI (2025))
0.0
Aime 25(MAA (Mathematical Association of America))
0.0

Scores par catégorie LLM Stats

(LLM Stats (zeroeval))
Creativity
90
Writing
90
Roleplay
80
Structured Output
80
Instruction Following
80
Code
80
Legal
70
Math
70
Reasoning
70
Language
70
Finance
70
General
70
Healthcare
70
Communication
70
Physics
50
Biology
50
Chemistry
50

Tarification

Prix d'entrée$0.1 / 1M tokens
Prix de sortie$0.3 / 1M tokens
Prix mixte (3:1)$0.15 / 1M tokens

Vitesse

Tokens/sec0.0
Délai du premier token0.00s
Temps de réponse0.00s

Classement des Prix par Fournisseur

Classement des Prix par Fournisseur

3 fournisseurs

Moins cher: OpenRouterPlus cher: Mistral
FournisseurEntréeSortie
1OpenRouterMoins cher
$0.05
$0.08
2Kilo Gateway
$0.05
$0.08
3MistralPRINCIPAL
$0.1
$0.3

Comparer les prix entre différents fournisseurs API pour ce modèle.

Sources externes