DataLearner logo
MA

Magistral-Small-2506

Reasoning model

Magistral-Small-2506

Release date: 2025-06-10Updated: 2025-06-10Views: 813
Live demoGitHubHugging FaceCompare
Parameters
24B
Context length
128K
Multilingual
Supported
Reasoning ability
3/5

Magistral-Small-2506 is an AI model published by MistralAI, released on 2025-06-10, for Reasoning model, with 24B parameters, and 128K context length, with a 70.68 score on AIME 2024.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Magistral-Small-2506

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
128K tokens
Max output length
64K tokens
Model type
Reasoning model
Modality (in / out)
Text → Text
Release date
2025-06-10
Model file size
No data
MoE architecture
No
Total params / Active params
24B / Not applicable
Knowledge cutoff
No data
Magistral-Small-2506

Open source & experience

Code license
Apache 2.0
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
N/A
Hugging Face
huggingface.co/mistralai/Magistral-Small-2506
Live demo
chat.mistral.ai/chat
Magistral-Small-2506

Official resources

Paper
Paper
DataLearnerAI blog
N/A
Magistral-Small-2506

API details

API speed
4/5
No public API pricing yet.
Magistral-Small-2506

Benchmark Results

Magistral-Small-2506 currently shows benchmark results led by AIME 2024 (38 / 62, score 70.68), LiveCodeBench (85 / 127, score 55.84), AIME2025 (79 / 106, score 62.76). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking

General Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Standard Mode
68.18
203 / 270

Math and Reasoning

2 evaluations
Benchmark / mode
Score
Rank/total
AIME 2024
Standard Mode
70.68
38 / 62
AIME2025
Standard Mode
62.76
79 / 106

Coding and Software Engineer

1 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Standard Mode
55.84
85 / 127
View benchmark analysisCompare with other models
Magistral-Small-2506

Publisher

Magistral-Small-2506

Model Overview

Magistral-Small-2506 is an AI model published by MistralAI, released on 2025-06-10, for Reasoning model, with 24B parameters, and 128K context length, with a 70.68 score on AIME 2024.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code