Magistral Small 24B

Overview

Mistral AI's first open reasoning model, built on Small 3.1 with RL-trained chain-of-thought. Hits 70.7% on AIME24 under Apache 2.0.

When to pick this model

• Math, science, and competition-style problem solving on local hardware
• Transparent reasoning where visible CoT helps debugging
• Reasoning workloads requiring a permissively licensed alternative to DeepSeek R1
• Multi-step planning agents with reasoning budgets under 40k tokens

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	14 GB
Q5_K_M	17 GB
Q8_0	26 GB
FP16 (no quantization)	48 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Published benchmark scores

Benchmark	Score
AIME 2024	70.7
MATH-500	90

Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.

Strengths

• First open Mistral reasoner with a real RL training pipeline
• AIME24 70.7% — competitive with much larger reasoners
• Apache 2.0 license
• Runs on a single 24GB GPU at Q4

Limitations

• Highly verbose in thinking mode — token costs add up
• Recommended effective context capped around 40k
• Trails DeepSeek R1 distills on hardest math benchmarks

Architecture & training

Architecture: Dense 24B · CoT reasoning · Small 3.1 base

Training: RL on reasoning.

Quick start

ollama run magistral:24b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.