Mistral Small 24B Instruct 2501: local hardware requirements and GPU compatibility
← All models

Local AI model

Mistral Small 24B Instruct 2501

by Mistral AI

3 of 22 reference machines run it entirely in accelerator memory at 8k context, and 10 more with CPU offload.

23.6B Parameters 32k Maximum context 2 Quantization Full Compatibility coverage
Check my PC

Overview

Publisher
Mistral AI
Family
Mistral
Series
Mistral Small
Task
Text generation
Architecture
dense · mistral
Parameters
23.6B
Maximum context
32,768 tokens
Licence
apache-2.0
Published
2025-01-28
Formats
gguf, safetensors
KV cache per token (f16)
163,840 B · standard attention, exact
Compatibility coverage
Full: verdicts computed
Upstream repository
mistralai/Mistral-Small-24B-Instruct-2501
Official page
https://mistral.ai/

Quantizations and file sizes

Quantization Runtime Weights Memory floor at 4k Identity
Q4_K_Mgguf ollamaollama run mistral-small:24b 13.3 GB 14.0 GB digest verified
Q8_0gguf ollamaollama run mistral-small:24b-instruct-2501-q8_0 23.3 GB 24.0 GB digest verified

Sizes come from the runtime registry or repository listing. The memory floor adds the KV cache for 4,096 tokens when that figure cannot over-count; it is an estimate, not a requirement.

Runtimes

Compatibility on reference machines

8k context, 32 GB system RAM, best catalogued quantization per machine.

Machine Memory Answer Known facts
GeForce RTX 3090 24GB 24.0 GB VRAM Runs fully in memory estimated · Q4_K_M Weights 13.3 GB vs 24.0 GB VRAM: fitKV cache at 8k: 1.25 GB
GeForce RTX 4090 24GB 24.0 GB VRAM Runs fully in memory estimated · Q4_K_M Weights 13.3 GB vs 24.0 GB VRAM: fitKV cache at 8k: 1.25 GB
GeForce RTX 5090 32GB 32.0 GB VRAM Runs fully in memory estimated · Q8_0 Weights 23.3 GB vs 32.0 GB VRAM: fitKV cache at 8k: 1.25 GB
GeForce RTX 3060 Laptop GPU 6GB 6.00 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 6.00 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 4060 8GB 8.00 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 8.00 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 3080 10GB 10.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 10.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 3060 12GB 12.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 12.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 4070 12GB 12.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 12.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 5070 12GB 12.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 12.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 4060 Ti 16GB 16.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 16.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 4080 16GB 16.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 16.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 5060 Ti 16GB 16.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 16.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
GeForce RTX 5080 16GB 16.0 GB VRAM Runs with CPU offload estimated · Q8_0 Weights 23.3 GB vs 16.0 GB VRAM: do not fitKV cache at 8k: 1.25 GB
Radeon RX 7800 XT 16GB 16.0 GB VRAM Potential: weights fit not verified · Q4_K_M Weights 13.3 GB vs 16.0 GB VRAM: fitKV cache at 8k: 1.25 GB
Radeon RX 9070 XT 16GB 16.0 GB VRAM Potential: weights fit not verified · Q4_K_M Weights 13.3 GB vs 16.0 GB VRAM: fitKV cache at 8k: 1.25 GB
Radeon RX 7900 XTX 24GB 24.0 GB VRAM Potential: weights fit not verified · Q4_K_M Weights 13.3 GB vs 24.0 GB VRAM: fitKV cache at 8k: 1.25 GB
Radeon RX 7600 8GB 8.00 GB VRAM Potential: weights need offload not verified · Q4_K_M Weights 13.3 GB vs 8.00 GB VRAM: do not fitKV cache at 8k: 1.25 GB
Intel Arc B580 12GB 12.0 GB VRAM Not enough data not verified · Q4_K_M Weights 13.3 GB vs 12.0 GB VRAM: do not fitKV cache at 8k: 1.25 GBruntime support on this machine: unknown
Intel Arc A770 16GB 16.0 GB VRAM Not enough data not verified · Q4_K_M Weights 13.3 GB vs 16.0 GB VRAM: fitKV cache at 8k: 1.25 GBruntime support on this machine: unknown
Mac mini M4 24GB 24.0 GB unified Not enough data not verified · Q4_K_M Weights 13.3 GB vs 24.0 GB unified: fitKV cache at 8k: 1.25 GB
Mac mini M4 Pro 24GB 24.0 GB unified Not enough data not verified · Q4_K_M Weights 13.3 GB vs 24.0 GB unified: fitKV cache at 8k: 1.25 GB
MacBook Pro 14-inch M4 Max 36GB 36.0 GB unified Not enough data not verified · Q4_K_M Weights 13.3 GB vs 36.0 GB unified: fitKV cache at 8k: 1.25 GB

Definitive answers ("runs", "does not fit") come from the calibrated engine or a physical run. "Potential" rows only compare the weights with memory: the rest of the requirement is not modelled, so they are not verdicts. Unknown is never a failure.

Hardware for this model

Reference GPUs with 24 GB of VRAM or more run this model entirely in memory at 8k context: GeForce RTX 3090 24GB, GeForce RTX 4090 24GB, GeForce RTX 5090 32GB.

  • Smallest catalogued artifact: 13.3 GB of weights (Q4_K_M).
  • A GPU with at least 24 GB of VRAM holds those weights.
Explore 24 GB VRAM GPUsExplore desktops with a 24 GB GPU
These links open a hardware category search, not a specific product recommendation. Affiliate link — I may earn a commission from qualifying purchases. These links never change a compatibility answer. They follow the memory the data defends, not commission.

Sources and evidence

0 measured, 37 sourced, 4 estimated and 5 unknown fields.

Catalogue record verified 2026-09-16.

What else can your PC run?

Pick your GPU and system RAM to see every catalogued model that fits, with tested, estimated and potential answers kept apart.

Check my PC