Mistral · 9B to 32B
Mistral Small 24B Instruct 2501
23.57 billion parameters across 40 layers. At four-bit precision the weights alone come to 13.25 GB, before any conversation is loaded. It uses grouped-query attention with 8 KV heads for 32 query heads, which cuts the cache to 25% of what full attention would need.
About Mistral Small 24B Instruct 2501
Mistral's 24B from January 2025, under Apache 2.0. At 4 bits it needs around 14 GB, which makes it a natural fit for a 16 GB card.
Mistral Small 24B Instruct 2501 is a dense model with 23.57B parameters, built from 40 layers with a hidden size of 5120. Its repository was published on 2025-01-28. The license listed on Hugging Face is apache-2.0. It uses grouped-query attention with 8 KV heads for 32 query heads, which cuts the cache to 25% of what full attention would need.
Memory needed, by quantization
At an 8k context window. Quality is our rough ranking of how much the compression costs you: anything at or above Q5 is hard to tell apart from the original in normal use.
| Quantization | File size | KV cache | Total VRAM | Quality | What it costs you |
|---|---|---|---|---|---|
| Q8_0 | 23.32 GB | 1.25 GB | 24.35 GB | 99% | Lossless in practice. Use it when the memory is there. |
| Q6_K | 18 GB | 1.25 GB | 19.18 GB | 98% | Very close to Q8 for two thirds of the size. |
| Q5_K_M | 15.61 GB | 1.25 GB | 16.86 GB | 96% | The quality-first choice when Q6 will not fit. |
| Q4_K_M | 13.25 GB | 1.25 GB | 14.57 GB | 93% | The default. Best size-to-quality ratio for local use. |
| IQ4_XS | 11.66 GB | 1.25 GB | 13.02 GB | 91% | Importance-matrix 4-bit. Q4_K_S quality, smaller file. |
| Q3_K_M | 10.73 GB | 1.25 GB | 12.11 GB | 86% | Degradation starts to show. A way to fit one size up. |
| IQ3_XXS | 8.4 GB | 1.25 GB | 9.85 GB | 79% | Aggressive. Only worth it on very large models. |
What a longer conversation costs
Same model at Q4_K_M, only the context window changes. The cache grows in a straight line with every token in the window.
| Context | KV cache | Total VRAM | Fits in 8 GB | Fits in 12 GB | Fits in 24 GB |
|---|---|---|---|---|---|
| 4k | 0.63 GB | 13.94 GB | no | no | yes |
| 8k | 1.25 GB | 14.57 GB | no | no | yes |
| 16k | 2.5 GB | 15.82 GB | no | no | yes |
| 32k | 5 GB | 18.32 GB | no | no | yes |
The "fits" columns allow for the roughly 0.8 GB Windows keeps for the desktop.
Which hardware runs Mistral Small 24B Instruct 2501
62 of 118 consumer devices run it at a quantization worth using, at an 8k context window. Another 27 can load it only by compressing the weights far enough to damage the model, marked with a warning below.
| Device | Memory | Verdict | Quantization | Used | Speed |
|---|---|---|---|---|---|
| GeForce RTX 5090 NVIDIA | 32 GB | Runs great | Q8_0 | 24.35 GB | 55.1 t/s |
| RTX PRO 6000 Blackwell NVIDIA | 96 GB | Runs great | Q8_0 | 24.35 GB | 55.1 t/s |
| RTX 6000 Ada Generation NVIDIA | 48 GB | Runs great | Q8_0 | 24.35 GB | 29.5 t/s |
| RTX A6000 NVIDIA | 48 GB | Runs well | Q8_0 | 24.35 GB | 23.6 t/s |
| Apple M3 Ultra 96GB Apple | 96 GB | Runs well | Q8_0 | 24.98 GB | 23.3 t/s |
| Apple M3 Ultra 256GB Apple | 256 GB | Runs well | Q8_0 | 24.98 GB | 23.3 t/s |
| Apple M3 Ultra 512GB Apple | 512 GB | Runs well | Q8_0 | 24.98 GB | 23.3 t/s |
| Apple M1 Ultra 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 22.7 t/s |
| Apple M1 Ultra 128GB Apple | 128 GB | Runs well | Q8_0 | 24.98 GB | 22.7 t/s |
| Apple M2 Ultra 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 22.7 t/s |
| Apple M2 Ultra 128GB Apple | 128 GB | Runs well | Q8_0 | 24.98 GB | 22.7 t/s |
| Apple M2 Ultra 192GB Apple | 192 GB | Runs well | Q8_0 | 24.98 GB | 22.7 t/s |
| Radeon PRO W7900 AMD | 48 GB | Runs well | Q8_0 | 24.35 GB | 22.5 t/s |
| GeForce RTX 4090 NVIDIA | 24 GB | Runs great | Q6_K | 19.18 GB | 39.6 t/s |
| GeForce RTX 3090 Ti NVIDIA | 24 GB | Runs great | Q6_K | 19.18 GB | 39.6 t/s |
| GeForce RTX 3090 NVIDIA | 24 GB | Runs great | Q6_K | 19.18 GB | 36.8 t/s |
| GeForce RTX 5090 Laptop NVIDIA | 24 GB | Runs great | Q6_K | 19.18 GB | 35.2 t/s |
| Radeon RX 7900 XTX AMD | 24 GB | Runs great | Q6_K | 19.18 GB | 32 t/s |
| RTX A5000 NVIDIA | 24 GB | Runs great | Q6_K | 19.18 GB | 30.2 t/s |
| Radeon RX 7900 XT AMD | 20 GB | Just fits | Q6_K | 19.18 GB | 26.6 t/s |
| Apple M4 Max 36GB Apple | 36 GB | Just fits | Q8_0 | 24.98 GB | 15.5 t/s |
| Apple M4 Max 48GB Apple | 48 GB | Runs well | Q8_0 | 24.98 GB | 15.5 t/s |
| Apple M4 Max 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 15.5 t/s |
| Apple M4 Max 128GB Apple | 128 GB | Runs well | Q8_0 | 24.98 GB | 15.5 t/s |
| GeForce RTX 5080 NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 50 t/s |
Show 35 more devices
| GeForce RTX 5070 Ti NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 46.7 t/s |
| GeForce RTX 5080 Laptop NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 40 t/s |
| GeForce RTX 4080 SUPER NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 38.3 t/s |
| GeForce RTX 4080 NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 37.3 t/s |
| GeForce RTX 4070 Ti SUPER NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 35 t/s |
| GeForce RTX 4090 Laptop NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 30 t/s |
| Radeon RX 9070 XT AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 28.5 t/s |
| Radeon RX 9070 AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 28.5 t/s |
| Radeon RX 7800 XT AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 27.6 t/s |
| Radeon RX 7900 GRE AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 25.4 t/s |
| GeForce RTX 5060 Ti 16GB NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 23.3 t/s |
| RTX A4000 NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 23.3 t/s |
| Radeon RX 6900 XT AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 22.6 t/s |
| Radeon RX 6800 AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 22.6 t/s |
| Arc A770 16GB Intel | 16 GB | Just fits | Q4_K_M | 14.57 GB | 22.5 t/s |
| Apple M1 Max 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M2 Max 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M2 Max 96GB Apple | 96 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M3 Max 36GB Apple | 36 GB | Just fits | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M3 Max 48GB Apple | 48 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M3 Max 64GB Apple | 64 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M3 Max 96GB Apple | 96 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| Apple M3 Max 128GB Apple | 128 GB | Runs well | Q8_0 | 24.98 GB | 11.4 t/s |
| GeForce RTX 4060 Ti 16GB NVIDIA | 16 GB | Just fits | Q4_K_M | 14.57 GB | 15 t/s |
| NVIDIA DGX Spark 128GB NVIDIA | 128 GB | Runs well | Q6_K | 19.69 GB | 10.7 t/s |
| Apple M4 Pro 48GB Apple | 48 GB | Runs well | Q5_K_M | 17.27 GB | 11.3 t/s |
| Apple M4 Pro 64GB Apple | 64 GB | Runs well | Q5_K_M | 17.27 GB | 11.3 t/s |
| Radeon RX 7600 XT AMD | 16 GB | Just fits | Q4_K_M | 14.57 GB | 12.7 t/s |
| Jetson AGX Orin 32GB NVIDIA | 32 GB | Runs well | Q4_K_M | 14.94 GB | 10.7 t/s |
| Jetson AGX Orin 64GB NVIDIA | 64 GB | Runs well | Q4_K_M | 14.94 GB | 10.7 t/s |
| Ryzen AI Max+ 395 32GB AMD | 32 GB | Runs well | Q4_K_M | 14.94 GB | 10.3 t/s |
| Ryzen AI Max+ 395 64GB AMD | 64 GB | Runs well | Q4_K_M | 14.94 GB | 10.3 t/s |
| Ryzen AI Max+ 395 96GB AMD | 96 GB | Runs well | Q4_K_M | 14.94 GB | 10.3 t/s |
| Ryzen AI Max+ 395 128GB AMD | 128 GB | Runs well | Q4_K_M | 14.94 GB | 10.3 t/s |
| Apple M1 Pro 32GB Apple | 32 GB | Runs well | IQ4_XS | 13.32 GB | 10.8 t/s |
How it compares
Other models in the same family or of a similar size, on a 12 GB card and a 24 GB card, each at the best quantization that fits.
| Model | Weights at 4 bits | GeForce RTX 3060 12GB | GeForce RTX 4090 |
|---|---|---|---|
| Mistral Small 24B Instruct 2501 | 13.3 GB | does not fit well | 39.6 t/s at Q6_K |
| Codestral 22B V0.1 | 12.5 GB | does not fit well | 39.9 t/s at Q6_K |
| Mistral Nemo Instruct 2407 | 6.9 GB | 26.2 t/s at Q6_K | 58.4 t/s at Q8_0 |
| Mistral 7B Instruct V0.3 | 4.1 GB | 32.9 t/s at Q8_0 | 92.2 t/s at Q8_0 |
| Gemma 4 31B Instruct | 17.6 GB | does not fit well | 35.5 t/s at Q4_K_M |
| Gemma 4 26B A4B Instruct | 14.5 GB | 37.6 t/s at Q4_K_M | 117 t/s at Q6_K |
What to buy to run Mistral Small 24B Instruct 2501
The cheapest hardware that runs it at a quantization worth using and a speed you would not resent, at an 8k context window.
Cheapest that works
Arc A770 16GB
$280 used
Q4_K_M · uses 14.57 GB of its 16 GB · about 22.5 tokens/s
Best value
Radeon RX 6800
$330 used
Q4_K_M · about 22.6 tokens/s · 512 GB/s
| Device | Price | Memory | Runs it at | Speed | |
|---|---|---|---|---|---|
| Arc A770 16GB Intel | $280 used | 16 GB | Q4_K_M | 22.5 t/s | Check price |
| Radeon RX 6800 AMD | $330 used | 16 GB | Q4_K_M | 22.6 t/s | Check price |
| GeForce RTX 5060 Ti 16GB NVIDIA | $430 | 16 GB | Q4_K_M | 23.3 t/s | Check price |
| Radeon RX 7800 XT AMD | $480 | 16 GB | Q4_K_M | 27.6 t/s | Check price |
| Radeon RX 9070 AMD | $560 | 16 GB | Q4_K_M | 28.5 t/s | Check price |
| Radeon RX 7900 XT AMD | $650 | 20 GB | Q6_K | 26.6 t/s | Check price |
| GeForce RTX 3090 NVIDIA | $700 used | 24 GB | Q6_K | 36.8 t/s | Check price |
| GeForce RTX 5070 Ti NVIDIA | $780 | 16 GB | Q4_K_M | 46.7 t/s | Check price |
Indicative prices reviewed 2026-09-01; used prices are marketplace typical. Options that cost more than a cheaper one with no more memory and no more speed are hidden. Price links are Amazon affiliate links. Change the standard, the context window or the budget in the buying tool.
Where it sits in the catalog
Source: mistralai/Mistral-Small-24B-Instruct-2501 on Hugging Face. Downloaded 45k times in the last month. Published 2025-01-28. Architecture figures are read from the repository's own configuration file, so they move when the model does.
Questions people ask about Mistral Small 24B Instruct 2501
How much VRAM does Mistral Small 24B Instruct 2501 need?
At Q4_K_M with an 8k context window, Mistral Small 24B Instruct 2501 needs about 14.57 GB: 12.88 GB of weights, 1.25 GB of KV cache and 0.44 GB of overhead. At the highest-quality file it needs about 24.35 GB.
What is the cheapest GPU that runs Mistral Small 24B Instruct 2501?
The smallest consumer device in our catalog that fits it is the GeForce RTX 5080 with 16 GB, running it at Q4_K_M at roughly 50 tokens per second.
How much does a longer context window add for Mistral Small 24B Instruct 2501?
On the memory side, a 32k window adds 5.0 GB of KV cache on top of the weights, and the cache grows in a straight line with every token. The longest window the model supports natively is 32k tokens.