---
title: "Mistral Small 3.1 24B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Mistral Small 3.1."
source: "https://rankedagi.com/models/mistral-small-3.1-24b"
---

# Mistral Small 3.1

| Field | Value |
| --- | --- |
| Organization | Mistral |
| License | Open Source |
| Size | 24B |
| Version | Latest |
| Released | 2025-03-17 |
| Context window | 128K |
| Last updated | 2026-04-20T06:30:02.027Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 26.4% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 22.6% | RankedAGI Agentic Score |  |
| RankedAGI Reasoning | reasoning | 33.0% | RankedAGI Reasoning Score |  |
| GPQA Diamond | reasoning | 46.0% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| HumanEval | coding | 88.4% | HumanEval coding benchmark |  |
| NYT Connections Extended | reasoning | 11.4% | NYT Connections Extended Version | https://github.com/lechmazur/nyt-connections/ |
| MMLU Pro | knowledge | 66.8% | Graduate-level visual problem-solving |  |
| MMLU | knowledge | 80.6% | Image Reasoning |  |
| MMMU | imaging | 62.8% | Multimodal Understanding College-level visual problem-solving |  |
| Hughes Hallucination Evaluation | safety | 5.6% | Computed using Vectara's Hallucination Evaluation Model, also known as HHEM. This evaluates how often an LLM introduces hallucinations when summarizing a document. | https://github.com/vectara/hallucination-leaderboard |
| RankedAGI Math | math | 41.4% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 34.8% | Overall RankedAGI score |  |
| GDPval-AA Elo | Agents | 344 | Office Tasks (Artificial Analysis) |  |

## Model Links

| Title | URL |
| --- | --- |
| Release | https://mistral.ai/news/mistral-small-3-1 |