Mistral AIMistralMistralForCausalLM

Mistral Small 24B Instruct 2501 — Hardware Requirements & GPU Compatibility

Chat

Mistral Small 24B Instruct is Mistral AI's January 2025 release targeting the mid-range parameter sweet spot. At 24 billion parameters it sits between lightweight 7B models and heavier 70B-class offerings, delivering strong instruction-following, reasoning, and coding performance without demanding top-tier hardware. This model fits comfortably on a single GPU with 16–24 GB of VRAM at common quantization levels, making it an attractive option for users with cards like the RTX 4090 or RTX 3090 who want a noticeable step up from 7B models. It strikes an appealing balance between quality and resource requirements for serious local use.

23.6B params · dense33K contextApache 2.0Jan 202594.0K downloads991 likes112.6K quant downloads

Your hardware

Verdicts below are for the reference RTX 4090. Set yours to personalise the page.

Detection reads your browser's GPU name with no data sent anywhere.

Runs well on a 24 GB GPU

Runs from 12 GB (IQ2_M · 16K)Best on 24 GB (Q4_K_M · ~44 tok/s · 32K)Macs from 16 GB (IQ2_M · ~16 tok/s)

Apache 2.0

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Mistral Small 24B Instruct 2501

One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.

Ways to runOther ways

each opens step-by-step

Run in cloud

Runs on RTX 3090 (24 GB) (6 GB headroom) · Q4_K_M

Generation speed: ~41 tok/s; generation speed
Cost per 1M output tokens: $1.34; per 1M output tokens

Deployon Vast.aiBest valuefrom $0.20/hr

Compare GPUs →

Run Mistral Small 24B Instruct 2501

17.8 GB VRAM at 16K context · fits the RTX 4090 (6.2 GB free)

Works on macOS, Windows and Linux.

Install Ollama (Linux)
$curl -fsSL https://ollama.com/install.sh | sh
Install Ollama (macOS / Windows)
Download and run the app; it installs the ollama command.
Download Ollama ↗
Run Q4_K_M
$ollama run mistral-small:24b
Official Ollama library build. For the exact Q4_K_M file from MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF: ollama run hf.co/MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF:Q4_K_M
Raise the context window
Type it inside the chat. For the API server instead, start it with OLLAMA_CONTEXT_LENGTH=16384 ollama serve.
$/set parameter num_ctx 16384
A larger context needs more memory than the weights alone.
Use the API
Ollama serves its API on http://localhost:11434.

Works on macOS, Windows and Linux.

Install LM Studio
Desktop app for macOS (Apple silicon), Windows and Linux.
lmstudio.ai ↗
Download the model
In the app press Ctrl/Cmd + Shift + M and paste https://huggingface.co/MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF@Q4_K_M
Open on Hugging Face ↗
Load and chat
Open the model loader, select the model and set Context Length to 16384. Leave GPU offload at maximum if it fits.
Start the local server (optional)
$lms server start

Works on macOS, Windows and Linux.

Install llama.cpp (macOS)
MacPorts, Nix and conda-forge also work.
$brew install llama.cpp
Install options ↗
Install llama.cpp (Windows)
$winget install llama.cpp
Install llama.cpp (Linux)
Download the prebuilt llama-*-bin-ubuntu-* archive for your GPU backend (CPU, Vulkan) from the releases page, unpack it and run the binaries from that folder. Homebrew (brew install llama.cpp) and conda-forge also work on Linux.
llama.cpp releases ↗
Start the server
Downloads the model on first run. The chat UI and an OpenAI-compatible API are at http://localhost:8080.
$llama-server -hf MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF:Q4_K_M -c 16384 -ngl 99 --port 8080

Works on macOS, Windows and Linux.

Option A: Docker Model Runner
Docker Desktop: Settings > AI > Enable Docker Model Runner. Docker Engine on Linux: install the docker-model-plugin package.
$docker model run hf.co/MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF:Q4_K_M
Model Runner docs ↗
The quant tag after the repo is how Hugging Face cards write it; if your Docker version rejects it, drop the tag and pick the quant by repo.
Option B: llama.cpp server image (NVIDIA)
Needs the NVIDIA Container Toolkit. Models are cached in the llama-cache volume. Open http://localhost:8080.
$docker run --gpus all -p 8080:8080 -v llama-cache:/root/.cache ghcr.io/ggml-org/llama.cpp:server-cuda -hf MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF:Q4_K_M -c 16384 -ngl 99 --host 0.0.0.0 --port 8080
Image tags (cuda13, vulkan, rocm) ↗
Cache path inside the image is not in the docs; remove the -v flag if the volume complains.

Linux only.

Install vLLM
Linux only; on Windows run it inside WSL2. NVIDIA GPU shown; AMD ROCm, Intel and CPU builds have their own install pages.
$uv venv --python 3.12 --seed
$source .venv/bin/activate
$uv pip install vllm --torch-backend=auto
vLLM installation ↗
Serve the model
OpenAI-compatible API at http://localhost:8000/v1.
$vllm serve mistralai/Mistral-Small-24B-Instruct-2501 --max-model-len 16384
Runs the original weights, not the Q4_K_M GGUF; needs ~52 GB of GPU memory.

macOS only.

MLX supports only macOS, so follow these steps for that setup.

Install mlx-lm
$pip install mlx-lm
Chat in the terminal
$mlx_lm.chat --model mlx-community/Mistral-Small-24B-Instruct-2501-4bit
Downloads the model on first run. The MLX repo has its own quantisation; it may not match the quant selected above.
Or start a local server
Listens on http://localhost:8080.
$mlx_lm.server --model mlx-community/Mistral-Small-24B-Instruct-2501-4bit

Works on macOS, Windows and Linux.

Install Jan
jan.ai ↗
Find the model
Open the Hub tab, paste this model ID in Search, and choose Q4_K_M if offered.
$MaziyarPanahi/Mistral-Small-24B-Instruct-2501-GGUF
Use the local API (optional)
Settings > Local API Server > Start Server. It listens on http://127.0.0.1:1337.

Rent GPUs hourly to run Q4_K_M right there. Pricing comes from partner marketplaces and varies.

Available GPU rentals for Q4_K_M, prioritized by value
GPU	Speed	$/1M tokens	Rent
RTX 3090 (24 GB)CheapestCheapest6 GB free	~41 tok/s	$1.34	Deploy on Vast.aifrom $0.20/hr (opens in new tab)	24 GB · 6 GB free · ~41 tok/s · $1.34 / 1M tokens
TITAN RTX (24 GB)6 GB free	~29 tok/s	$2.04	Deploy on Vast.aifrom $0.22/hr (opens in new tab)	24 GB · 6 GB free · ~29 tok/s · $2.04 / 1M tokens
RTX A5000 (24 GB)6 GB free	~34 tok/s	$1.89	Deploy on Vast.aifrom $0.23/hr (opens in new tab)	24 GB · 6 GB free · ~34 tok/s · $1.89 / 1M tokens
RTX PRO 4000 (24 GB)6 GB free	~29 tok/s	$2.54	Deploy on Vast.aifrom $0.27/hr (opens in new tab)	24 GB · 6 GB free · ~29 tok/s · $2.54 / 1M tokens
RTX PRO 4500 (32 GB)14 GB free	~39 tok/s	$2.41	Deploy on RunPodfrom $0.34/hr (opens in new tab)	32 GB · 14 GB free · ~39 tok/s · $2.41 / 1M tokens
RTX 4090 (24 GB)6 GB free	~44 tok/s	$2.14	Deploy on RunPodfrom $0.34/hr (opens in new tab)	24 GB · 6 GB free · ~44 tok/s · $2.14 / 1M tokens
A40 (48 GB)30 GB free	~30 tok/s	$3.20	Deploy on RunPodfrom $0.35/hr (opens in new tab)	48 GB · 30 GB free · ~30 tok/s · $3.20 / 1M tokens
L4 (24 GB)6 GB free	~13 tok/s	$9.33	Deploy on RunPodfrom $0.44/hr (opens in new tab)	24 GB · 6 GB free · ~13 tok/s · $9.33 / 1M tokens

Compare GPUs →

How Much VRAM Does Mistral Small 24B Instruct 2501 Need?

Select a quantization level; the listed hardware adheres to it.

VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
7.2 GB	13.5 GB	—	6.5 GB	2-bit · i-quant extreme compression — Importance-weighted 2-bit, extreme compression — significant quality loss	Fits: 32K context · ~91 tok/s
7.8 GB	14.1 GB	—	7.2 GB	2-bit · i-quant extra small — Importance-weighted 2-bit, extra small	Fits: 32K context · ~84 tok/s
8.1 GB	14.4 GB	—	7.5 GB	2-bit · i-quant small — Importance-weighted 2-bit, small	Fits: 32K context · ~81 tok/s
8.7 GB	15 GB	—	8.1 GB	2-bit · i-quant medium — Importance-weighted 2-bit, medium	Fits: 32K context · ~76 tok/s
9.8 GB	16.1 GB	—	9.3 GB	3-bit · i-quant — Importance-weighted 3-bit	Fits: 32K context · ~67 tok/s
10.2 GB	16.4 GB	—	8.3 GB	2-bit · small K-quant — 2-bit small K-quant	Fits: 32K context · ~65 tok/s
10.4 GB	16.7 GB	—	9.9 GB	3-bit · i-quant extra small — Importance-weighted 3-bit, extra small	Fits: 32K context · ~63 tok/s
10.7 GB	17 GB	—	8.9 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 32K context · ~61 tok/s
10.7 GB	17 GB	—	10.4 GB	3-bit · i-quant small — Importance-weighted 3-bit, small	Fits: 32K context · ~61 tok/s
11.0 GB	17.3 GB	—	10.4 GB	3-bit small — 3-bit small quantization	Fits: 32K context · ~59 tok/s
11.3 GB	17.6 GB	—	10.7 GB	3-bit · i-quant medium — Importance-weighted 3-bit, medium	Fits: 32K context · ~58 tok/s
12.2 GB	18.5 GB	—	11.5 GB	3-bit medium — 3-bit medium quantization	Fits: 32K context · ~54 tok/s
12.5 GB	18.8 GB	—	13.5 GB	4-bit · legacy — 4-bit legacy quantization	Fits: 32K context · ~52 tok/s
12.8 GB	19.1 GB	—	12.4 GB	3-bit large — 3-bit large quantization	Fits: 32K context · ~51 tok/s
13.4 GB	19.7 GB	—	12.8 GB	4-bit · i-quant compact — Importance-weighted 4-bit, compact	Fits: 32K context · ~49 tok/s
14.0 GB	20.3 GB	—	13.5 GB	4-bit small — 4-bit small quantization	Fits: 32K context · ~47 tok/s
14.0 GB	20.3 GB	—	14.9 GB	4-bit · legacy+ — 4-bit legacy quantization with offset	Fits: 32K context · ~47 tok/s
14.0 GB	20.3 GB	—	13.5 GB	4-bit · i-quant non-linear — Importance-weighted 4-bit, non-linear	Fits: 32K context · ~47 tok/s
14.9 GB	21.2 GB	—	14.3 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 32K context · ~44 tok/s
15.2 GB	21.5 GB	—	14.8 GB	4-bit large — 4-bit large quantization	Fits: 32K context · ~43 tok/s
15.4 GB	21.7 GB	—	16.4 GB	5-bit · legacy — 5-bit legacy quantization	Fits: 32K context · ~42 tok/s
16.9 GB	23.2 GB	—	16.3 GB	5-bit small — 5-bit small quantization	Fits: 32K context · ~39 tok/s
16.9 GB	23.2 GB	—	17.7 GB	5-bit · legacy+ — 5-bit legacy quantization with offset	Fits: 32K context · ~39 tok/s
17.5 GB	23.8 GB	—	16.8 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 32K context · ~37 tok/s
17.8 GB	24.1 GB	—	17.2 GB	5-bit large — 5-bit large quantization	Fits: 16K context · ~37 tok/s
20.2 GB	26.5 GB	—	19.3 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: 16K context · ~33 tok/s
24.3 GB	30.6 GB	—	25.1 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Does not fit: needs 27.2 GB
47.9 GB	54.2 GB	—	47.2 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Does not fit: needs 50.8 GB

Selected Q4_K_M

4-bit medium quantization

17.8GBat 16K context

Weights: 14.1 GB
KV cache: 3.4 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 6.2 GB free · ~44 tok/s

≥50 tok/s needs ≥1.14 TB/s · RTX 5090 ~78

Which Hardware Can Run Mistral Small 24B Instruct 2501?

Q4_K_M · 14.9 GB

Mistral Small 24B Instruct 2501 (Q4_K_M) requires 14.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 20+ GB is recommended. Using the full 33K context window can add up to 6.3 GB, bringing total usage to 21.1 GB. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080. 122 devices with unified memory can run Mistral Small 24B Instruct 2501, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Mini M5 Pro (24 GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Doesn't fit — smallest quant needs 10.1 GB8 GB tier →

12 GB · RTX 3060 12GBIQ2_M · 11.6 GB · ~27 tok/s · 16K11.6~2716K12 GB tier →

16 GB · RTX 5060 Ti 16GBQ4_K_M · 17.8 GB · ~20 tok/s · 4Ktight17.8~204K16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 17.8 GB · ~44 tok/s · 32K17.8~4432K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 17.8 GB · ~6 tok/s · 32K17.8~632KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 17.8 GB · ~26 tok/s · 32K17.8~2632KMacs →

How good is Mistral Small 24B Instruct 2501?

llmrun Score blends public benchmark boards into one number; 100 = the average open model.

Not enough data yet for an llmrun Score.

A Score needs results on enough independent boards. How the Composite Score is built

Benchmarks

GPQA Diamond
Epoch AI · Jan 2025Independent
47.3%
#63/100 open
Independent
LMArena Text
LMArena · Oct 2026Arena
1274.4
#122/188 open
Arena
AIME 2024/2025
Epoch AI · Jan 2025Independent
6.7%
#68/88 open
Independent
MATH Level 5
Epoch AI · Jan 2025Independent
44.8%
#18/39 open
Independent

All 4 Mistral Small 24B Instruct 2501 benchmarks

Alternatives to Mistral Small 24B Instruct 2501

Mistral family · 35 models →

Similar models

3tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
LFM2.5 24B A2BLiquid AI	—	15 GB	~221	same size
Ariarhymes-ai	—	20.2 GB	~168	same size
EuroLLM 22B Instruct 2512utter-project	—	17.5 GB	~46	same size

Fine-tunes & variants(7)community builds of Mistral Small 24B Instruct 2501

Specifications

Publisher: Mistral AI
Family: Mistral
Parameters: 23.6B
Architecture: MistralForCausalLM
Context Length: 32,768 tokens
Vocabulary Size: 131,072
Release Date: 2025-01-28
License: Apache 2.0

Resources

Frequently Asked Questions

How much VRAM does Mistral Small 24B Instruct 2501 need?

Mistral Small 24B Instruct 2501 requires 14.9 GB of VRAM at Q4_K_M, or 47.9 GB at BF16. Full 33K context adds up to 6.3 GB (21.1 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 23.6B × 4.8 bits ÷ 8 = 14.1 GB

KV Cache + Overhead ≈ 0.8 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

KV Cache + Overhead ≈ 7 GB (at full 33K context)

VRAM usage by quantization

Q4_K_M

14.9 GB

Q4_K_M + full context

21.1 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 4090 run Mistral Small 24B Instruct 2501?

Yes, at Q6_K (20.2 GB) or lower. Higher quantizations like Q8_0 (24.3 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.

What's the best quantization for Mistral Small 24B Instruct 2501?

For Mistral Small 24B Instruct 2501, Q4_K_M (14.9 GB) offers the best balance of quality and VRAM usage. Q4_K_L (15.2 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 7.2 GB.

VRAM requirement by quantization

IQ2_XXS

7.2 GB~53%

Q2_K

10.7 GB~75%

IQ4_XS

13.4 GB~87%

Q4_K_M ★

14.9 GB~89%

Q5_0

15.4 GB~90%

BF16

47.9 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Mistral Small 24B Instruct 2501 on a Mac?

Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Mistral Small 24B Instruct 2501, the smallest at IQ2_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 14.9 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M5 Pro (24 GB).

Can I run Mistral Small 24B Instruct 2501 locally?

Yes — Mistral Small 24B Instruct 2501 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Mistral Small 24B Instruct 2501?

At Q4_K_M, Mistral Small 24B Instruct 2501 can reach ~323 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~44 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = (bandwidth GB/s ÷ model GB) × efficiency

Example: NVIDIA B200 → 8000 ÷ 14.86 × 0.65 = ~350 tok/s

Estimated speed at Q4_K_M (14.9 GB)

NVIDIA B200

~350 tok/s

NVIDIA GeForce RTX 4090

~44 tok/s

NVIDIA B300

~350 tok/s

AMD Instinct MI350X

~323 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of Mistral Small 24B Instruct 2501?

At Q4_K_M, the download is about 14.14 GB. The full-precision BF16 version is 47.14 GB. The smallest option (IQ2_XXS) is 6.48 GB.

Which GPUs can run Mistral Small 24B Instruct 2501?

26 consumer GPUs can run Mistral Small 24B Instruct 2501 at Q4_K_M. Top options include AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090, NVIDIA GeForce RTX 3090 Ti, AMD Radeon RX 6800. 7 GPUs have plenty of headroom for comfortable inference.

Which devices can run Mistral Small 24B Instruct 2501?

122 devices with unified memory can run Mistral Small 24B Instruct 2501 at Q4_K_M (14.9 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.