ZMIME
Mistral AI · Released Jun 10, 2025

Magistral Small

Open Mistral reasoning model for transparent step-by-step problem solving.

  • Open weights
  • Reasoning
  • Tool calling
Capability
133.2ECI · #120 of 148
Input
$0.50per 1M tokens
Output
$1.50per 1M tokens
Context
131K8K max output
Pricing

What it costs

Input
$0.50per million tokens
Output
$1.50per million tokens
Blended (3:1)
$0.75cheaper than 50% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$1.75
Team assistant25M input + 5M output tokens$20.00
Production app250M input + 50M output tokens$200.00

Median of 1 paid API provider on models.dev.

Capability

Independent benchmarks

133.2Epoch Capabilities Index
#120 of 148 scored models
88Median 146167

The shaded band is Epoch AI’s confidence range (128.9–134.4); the tick marks the median scored model.

  • GPQA DiamondGraduate-level science questions56.1%
  • OTIS Mock AIME 2024–2025Competition mathematics30.0%

Scores from Epoch AI, run independently of Mistral AI.

Specifications

The details

Lab
Mistral AI
Released
Jun 10, 2025
Context window
131,072 tokens
Max output
8,192 tokens
Inputs
Text
Output
Text
Reasoning
Always on
Tool calling
Yes
Structured output
No
Availability
1 API providerlisted on models.dev
Same lab

More from Mistral AI

Questions

About Magistral Small

How much does Magistral Small cost?

Magistral Small costs $0.50 input / $1.50 output per million tokens (median across 1 API provider). At a 3:1 input-to-output mix that is $0.75 per million tokens, cheaper than 50% of the 360 priced models we track.

What is the context window of Magistral Small?

Magistral Small accepts up to 131,072 tokens per request and can write up to 8,192 tokens in one response.

How good is Magistral Small?

Epoch AI gives Magistral Small a Capabilities Index score of 133.2 (likely range 128.9–134.4), ranking it #120 of 148 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 56.1%, OTIS Mock AIME 2024–2025 30.0%.

Is Magistral Small open source?

Yes. Mistral AI publishes the weights under the Apache 2.0 licence, so it can be downloaded and self-hosted.

What inputs does Magistral Small support?

Magistral Small accepts text and replies in text. It is a reasoning model, supports tool calling.

When was Magistral Small released?

Mistral AI released Magistral Small on Jun 10, 2025.

What are the best alternatives to Magistral Small?

The closest current models from other labs on capability, price and release date are GPT-4.1 mini (OpenAI, ECI 135.0, $0.40 / $1.60), Llama 4 Maverick 17B Instruct (Meta, ECI 132.2, $0.321 / $0.91), DeepSeek-V3 (DeepSeek, ECI 132.3, $0.32 / $1.10) and QwQ 32B (Alibaba (Qwen), ECI 137.6, $0.66 / $1.00).