Skip to content

Magistral Small 24B

by Mistral AI · mistral family

24B

parameters

text-generation code-generation reasoning math multilingual

Magistral Small 24B is Mistral AI's reasoning-focused model designed for complex problem-solving with transparent chain-of-thought capabilities. It delivers strong performance on mathematical reasoning, code generation, and multilingual tasks. With its 131K context window and efficient 24B parameter count, Magistral Small strikes a balance between reasoning capability and resource requirements, making it accessible for local deployment on mid-range hardware.

Quick Start with Ollama

ollama run q4_K_M
Resources Ollama Hugging Face Official Page
Creator Mistral AI
Parameters 24B
Architecture transformer-decoder
Context 128K tokens
Released Jun 18, 2025
License Apache 2.0
Ollama magistral

Quantization Options

Format File Size VRAM Required Quality Ollama Tag
Q4_K_M rec 14 GB 17 GB q4_K_M
Q8_0 25 GB 28 GB q8_0
F16 48 GB 52 GB fp16

Compatible Hardware

Q4_K_M requires 17 GB VRAM

Compatible Hardware

HardwareVRAMTypeFitEst. Speed
Mac Studio M4 Ultra 512GB512 GBmacRuns~48 tok/s
Mac Pro M2 Ultra 192GB192 GBmacRuns~47 tok/s
Mac Studio M4 Ultra 192GB192 GBmacRuns~48 tok/s
Mac Studio M4 Max 128GB128 GBmacRuns~32 tok/s
MacBook Pro M4 Max 128GB128 GBmacRuns~32 tok/s
MacBook Pro M5 Max 128GB128 GBmacRuns~32 tok/s
NVIDIA RTX PRO 6000 Blackwell96 GBgpuRuns~113 tok/s
MacBook Pro M3 Max 96GB96 GBmacRuns~24 tok/s
Mac mini M4 Pro 64GB64 GBmacRuns~16 tok/s
Mac Studio M4 Max 64GB64 GBmacRuns~32 tok/s
MacBook Pro M4 Max 64GB64 GBmacRuns~32 tok/s
MacBook Pro M5 Max 64GB64 GBmacRuns~32 tok/s
NVIDIA RTX 6000 Ada Generation48 GBgpuRuns~56 tok/s
NVIDIA RTX A600048 GBgpuRuns~45 tok/s
NVIDIA RTX PRO 5000 Blackwell48 GBgpuRuns~56 tok/s
Mac mini M4 Pro 48GB48 GBmacRuns~16 tok/s
MacBook Pro M3 Max 48GB48 GBmacRuns~24 tok/s
MacBook Pro M4 Pro 48GB48 GBmacRuns~16 tok/s
MacBook Pro M4 Max 48GB48 GBmacRuns~32 tok/s
MacBook Pro M5 Max 48GB48 GBmacRuns~24 tok/s
MacBook Pro M5 Pro 48GB48 GBmacRuns~16 tok/s
Mac Studio M4 Max 36GB36 GBmacRuns~32 tok/s
MacBook Pro M3 Pro 36GB36 GBmacRuns~9 tok/s
MacBook Pro M5 Max 36GB36 GBmacRuns~24 tok/s
NVIDIA RTX 5000 Ada Generation32 GBgpuRuns~42 tok/s
NVIDIA GeForce RTX 509032 GBgpuRuns~105 tok/s
iMac M4 32GB32 GBmacRuns~7 tok/s
Mac mini M4 32GB32 GBmacRuns~7 tok/s
MacBook Air M4 32GB32 GBmacRuns~7 tok/s
MacBook Air M5 32GB32 GBmacRuns~7 tok/s
MacBook Pro M5 32GB32 GBmacRuns~7 tok/s
AMD Radeon RX 7900 XTX24 GBgpuRuns~56 tok/s
NVIDIA GeForce RTX 3090 Ti24 GBgpuRuns~59 tok/s
NVIDIA GeForce RTX 309024 GBgpuRuns~55 tok/s
NVIDIA GeForce RTX 409024 GBgpuRuns~59 tok/s
NVIDIA RTX A500024 GBgpuRuns~45 tok/s
iMac M3 24GB24 GBmacRuns~6 tok/s
Mac mini M2 24GB24 GBmacRuns~6 tok/s
Mac mini M4 Pro 24GB24 GBmacRuns~16 tok/s
MacBook Air M2 24GB24 GBmacRuns~6 tok/s
MacBook Air M4 24GB24 GBmacRuns~7 tok/s
MacBook Air M5 24GB24 GBmacRuns~7 tok/s
MacBook Pro M4 Pro 24GB24 GBmacRuns~16 tok/s
MacBook Pro M5 24GB24 GBmacRuns~7 tok/s
MacBook Pro M5 Pro 24GB24 GBmacRuns~16 tok/s
AMD Radeon RX 7900 XT20 GBgpuRuns~47 tok/s
NVIDIA RTX 4000 Ada Generation20 GBgpuRuns~21 tok/s
MacBook Pro M3 Pro 18GB18 GBmacRuns (tight)~9 tok/s
AMD Radeon RX 6800 XT16 GBgpuCPU Offload~9 tok/s
AMD Radeon RX 6900 XT16 GBgpuCPU Offload~9 tok/s
AMD Radeon RX 7800 XT16 GBgpuCPU Offload~11 tok/s
AMD Radeon RX 9060 XT 16GB16 GBgpuCPU Offload~10 tok/s
AMD Radeon RX 9070 XT16 GBgpuCPU Offload~11 tok/s
AMD Radeon RX 907016 GBgpuCPU Offload~10 tok/s
Intel Arc A77016 GBgpuCPU Offload~10 tok/s
NVIDIA GeForce RTX 4060 Ti 16GB16 GBgpuCPU Offload~5 tok/s
NVIDIA GeForce RTX 4070 Ti Super16 GBgpuCPU Offload~12 tok/s
NVIDIA GeForce RTX 4080 Super16 GBgpuCPU Offload~13 tok/s
NVIDIA GeForce RTX 408016 GBgpuCPU Offload~13 tok/s
NVIDIA GeForce RTX 5060 Ti 16GB16 GBgpuCPU Offload~8 tok/s
NVIDIA GeForce RTX 5070 Ti16 GBgpuCPU Offload~16 tok/s
NVIDIA GeForce RTX 508016 GBgpuCPU Offload~17 tok/s
NVIDIA RTX A400016 GBgpuCPU Offload~8 tok/s
iMac M1 16GB16 GBmacCPU Offload~1 tok/s
iMac M4 16GB16 GBmacCPU Offload~2 tok/s
Mac mini M1 16GB16 GBmacCPU Offload~1 tok/s
Mac mini M4 16GB16 GBmacCPU Offload~2 tok/s
MacBook Air M2 16GB16 GBmacCPU Offload~2 tok/s
MacBook Air M3 16GB16 GBmacCPU Offload~2 tok/s
MacBook Air M4 16GB16 GBmacCPU Offload~2 tok/s
MacBook Air M5 16GB16 GBmacCPU Offload~2 tok/s
MacBook Pro M1 16GB16 GBmacCPU Offload~1 tok/s
MacBook Pro M2 Pro 16GB16 GBmacCPU Offload~4 tok/s
MacBook Pro M5 16GB16 GBmacCPU Offload~2 tok/s
AMD Radeon RX 6700 XT12 GBgpuCPU Offload~7 tok/s
AMD Radeon RX 7700 XT12 GBgpuCPU Offload~8 tok/s
Intel Arc B58012 GBgpuCPU Offload~8 tok/s
NVIDIA GeForce RTX 3060 12GB12 GBgpuCPU Offload~6 tok/s
NVIDIA GeForce RTX 3080 12GB12 GBgpuCPU Offload~16 tok/s
NVIDIA GeForce RTX 4070 Super12 GBgpuCPU Offload~9 tok/s
NVIDIA GeForce RTX 4070 Ti12 GBgpuCPU Offload~9 tok/s
NVIDIA GeForce RTX 407012 GBgpuCPU Offload~9 tok/s
NVIDIA GeForce RTX 507012 GBgpuCPU Offload~12 tok/s
24 hardware device(s) cannot run this model at Q4_K_M.

Benchmark Scores

77.0
mmlu