LLMs · Mistral AI

Magistral Small 1.2 (24B)

Name: Magistral Small 1.2 (24B)
Author: Mistral AI

Parameters: 24B
Architecture: Dense
Max context: 128K tokens
License: apache-2.0
Released: 2025-09

How to run it

Quickest · Ollama

One command downloads a ready-to-run version and starts a chat.

ollama run magistral:24b

Get Ollama at ollama.com/download (opens in a new tab) · model page (opens in a new tab)

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

huggingface.co/mistralai/Magistral-Small-2509-GGUF (opens in a new tab)

ollama run hf.co/mistralai/Magistral-Small-2509-GGUF

Original weights · Hugging Face

Full-precision files from Mistral AI, for fine-tuning, vLLM, Transformers.

huggingface.co/mistralai/Magistral-Small-2509 (opens in a new tab)

No sign-in needed.

Official repo · GitHub

Reference code, inference scripts and docs from Mistral AI. The weights themselves download from Hugging Face.

github.com/mistralai/mistral-inference (opens in a new tab)

git clone https://github.com/mistralai/mistral-inference

Mistral AI’s official inference code.

Links checked 5 days ago against Hugging Face, Ollama and GitHub.

Memory needed

Weights plus KV cache and runtime overhead. Leave a little extra free for the operating system and other apps.

Quantization	Quality	Weights	Total at 8K context	Total at 32K context	Download
`Q2_K`	Low	9.6 GB	11.5 GB	15.6 GB	GGUF files (opens in a new tab)
`Q3_K_M`	Low	11.8 GB	13.7 GB	17.7 GB	GGUF files (opens in a new tab)
`Q4_K_M`	Good	14.6 GB	16.6 GB	20.6 GB	GGUF (opens in a new tab)
`Q5_K_M`	High	17.0 GB	19.0 GB	23.0 GB	GGUF (opens in a new tab)
`Q6_K`	High	19.7 GB	21.6 GB	25.7 GB	GGUF files (opens in a new tab)
`Q8_0`	Very high	25.5 GB	27.4 GB	31.4 GB	GGUF (opens in a new tab)
`F16`	Full	48.0 GB	50.0 GB	54.0 GB	GGUF (opens in a new tab) Original weights (opens in a new tab)

Best hardware for Magistral Small 1.2 (24B)

The least expensive GPU in our catalog that runs it without offloading is the AMD Radeon RX 7600 XT (launch price $329) at Q3_K_M.

Grade · points	Hardware	Quant	Memory	Speed	Fit
87 ptsExcellent (S) · 87 of 100 points: Model quality 32.1/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 5090	`Q6_K`	23.2 GB / 32.0 GB	63.6 tok/s	Runs well
86 ptsExcellent (S) · 86 of 100 points: Model quality 31.4/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 4090	`Q5_K_M`	20.2 GB / 24.0 GB	42.6 tok/s	Runs well
86 ptsExcellent (S) · 86 of 100 points: Model quality 31.4/45, Memory fit 25.0/25, Speed 14.9/15, Memory use 15.0/15	NVIDIA GeForce RTX 3090	`Q5_K_M`	20.2 GB / 24.0 GB	39.5 tok/s	Runs well
85 ptsExcellent (S) · 85 of 100 points: Model quality 30.8/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 15.0/15	AMD Radeon RX 7900 XTX	`Q4_K_M`	17.8 GB / 24.0 GB	42.3 tok/s	Runs well
81 ptsGreat (A) · 81 of 100 points: Model quality 27.2/45, Memory fit 24.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 4070 Ti SUPER	`Q3_K_M`	14.5 GB / 16.0 GB	44.4 tok/s	Tight fit
81 ptsGreat (A) · 81 of 100 points: Model quality 27.2/45, Memory fit 24.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 4080 SUPER	`Q3_K_M`	14.5 GB / 16.0 GB	48.6 tok/s	Tight fit
81 ptsGreat (A) · 81 of 100 points: Model quality 27.2/45, Memory fit 24.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 5070 Ti	`Q3_K_M`	14.5 GB / 16.0 GB	59.2 tok/s	Tight fit
81 ptsGreat (A) · 81 of 100 points: Model quality 27.2/45, Memory fit 24.0/25, Speed 15.0/15, Memory use 15.0/15	NVIDIA GeForce RTX 5080	`Q3_K_M`	14.5 GB / 16.0 GB	63.5 tok/s	Tight fit
80 ptsGreat (A) · 80 of 100 points: Model quality 27.2/45, Memory fit 24.0/25, Speed 14.3/15, Memory use 15.0/15	AMD Radeon RX 9070 XT	`Q3_K_M`	14.5 GB / 16.0 GB	36.4 tok/s	Tight fit
80 ptsGreat (A) · 80 of 100 points: Model quality 32.4/45, Memory fit 25.0/25, Speed 15.0/15, Memory use 7.7/15	NVIDIA RTX PRO 6000 Blackwell	`Q8_0`	32.2 GB / 96.0 GB	46.6 tok/s	Runs well

Check it on your own hardware

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.