Local LLM model page

Magistral (24B)

Mistral efficient reasoning model. Strong chain-of-thought at 24B. 477K downloads.

Find the best model for my hardware Browse all 183 LLMs

Parameters

24B

Minimum RAM

20 GB

Model size

14 GB

Quantization

Q4_K_M

Can Magistral (24B) run locally?

Magistral (24B) is best suited for power-user machines with 32 GB RAM. LocalClaw recommends Q4_K_M as the default quantization, with at least 20 GB RAM.

Search term for LM Studio or compatible runtimes: magistral-24b

Hugging Face repository: mistralai/Magistral-24B-GGUF

reasoningpower

Strengths

Efficient reasoning model
Apache 2.0
Strong chain-of-thought
477K downloads

Limitations

Needs 20GB+ RAM
Reasoning overhead
New — less community support

Best use cases

Complex reasoning
Math
Strategic planning
Analysis

Benchmarks

Speed: 5/10

Quality: 8/10

Coding: 7/10

Reasoning: 9/10

Technical details

Developer: Mistral AI

License: Apache 2.0

Context window: 131,072 tokens

Architecture: Transformer with reasoning-enhanced training

Released: 2025-07