All times Eastern (ET)Provider pages checked daily at
ProSign inRSS

Mistral Small 3.0

API ID mistral-small-2501

Mistral Small 3.0 was a Mistral model, released January 2025. It had a 32,000-token context window. It took PDF input and supported function calling and structured outputs.

A small efficient and powerfull 24B open model for personal or commercial use.

Mistral's description
Status
Retired
Released
January 2025
Shutdown
Nov 30, 2025

mistral-small-2501 stopped serving on November 30, 2025. Recommended replacement: mistral-small-2603.

How to migrate to Mistral Small 4

Benchmarks

Scores as each source published them
or

Comparing with Mistral Small 2.0, the previous model in this line. A score is compared only with the same benchmark from the same source, at the same setting where both have it. Full comparison: prices, limits, features

Arena

ArenaMistral Small 3.0Mistral Small 2.0ChangeRankDateSource
Text
By category
  • Math 1261 rank 288 vs not rated
  • Creative writing 1227 rank 309 vs not rated
  • Instruction following 1255 rank 307 vs not rated
  • Hard prompts 1285 rank 298 vs not rated
  • Long prompts 1280 rank 305 vs not rated
  • Multi-turn 1250 rank 309 vs not rated
  • Expert 1259 rank 298 vs not rated
1274 ±6not rated307 of 415
14,681 votes		Oct 13, 2026source
Text: Coding1312 ±12not rated298 of 410
2,083 votes		Oct 13, 2026source

Rated as mistral-small-24b-instruct-2501. Ratings come from people voting between two anonymous answers; Text is style-controlled, as on arena.ai. Data: Arena (arena.ai), CC BY 4.0.

Epoch AI

BenchmarkMistral Small 3.0Mistral Small 2.0ChangeSettingRunSource
GPQA Diamond47.3%not runAug 30, 2026source
MATH Level 544.8%not runJan 30, 2025source
OTIS Mock AIME 2024-20256.7%not runAug 30, 2026source

Run independently by Epoch AI. Data: Epoch AI, Capabilities & benchmarking (epoch.ai/benchmarks), CC BY.

Features

Compare models
FeatureValueSinceSource
Structured outputsThe model can be held to a JSON schema you supply, so every reply parses without retries or cleanup.Supportedin place when tracking began (October 6, 2026)source
Function callingAlso called tool use. You describe functions; the model replies with a function name and arguments for your code to run, then uses the result.Supportedin place when tracking began (October 6, 2026)source
Batch APISend many requests as one job that runs in the background, usually finished within 24 hours, at a lower price than live requests.Supportedin place when tracking began (October 6, 2026)source
PDF inputPDF files can go in the prompt directly, without converting them to text first.Supportedin place when tracking began (October 6, 2026)source
Context windowThe most the model can work with in one request, in tokens: your prompt, conversation history and files. A token is about three-quarters of an English word.32,000 tokensin place when tracking began (October 6, 2026)source

Not published by Mistral: Prompt caching, Image input, Reasoning / thinking, Fine-tuning, Built-in web search, temperature parameter, top_p parameter, logprobs, Max output tokens.

Price history

USD per 1M tokens, standard tier
$0.00$0.10$0.20$0.30$0.40Oct 2026In $0.10Out $0.30Oct 5, 2026 to Oct 15, 2026: Input $0.10, Output $0.30
$0.00$0.10$0.20$0.30$0.40Oct 2026In $0.10Out $0.30Oct 5, 2026 to Oct 15, 2026: Input $0.10, Output $0.30

We started recording this model's prices on Oct 5, 2026. The history grows with each daily check.

Show as a table
FromTypePrice
Oct 5, 2026Input$0.10
Oct 5, 2026Output$0.30

Prices

USD per 1M tokens, every type and band, by tier
TypePrice $/1MIn effect
Standard
input$0.10
output$0.30

Changelog

Changes we detected on Mistral's pages

No changes recorded yet.

Last checked · Sources: Mistral models