Skip to content

Mistral Small 3.2 24B VRAM requirements

How much VRAM Mistral Small 3.2 24B needs
4-bit weights: 13.6 GBWith 8K context: 15.7 GBWith 32K context: 19.4 GB8-bit, 8K context: 25.9 GB
Parameters
24B
Longest context
128K
Released
Jun 2025
License
Apache 2.0
Strongest at
Chat, Coding, Agents

How much VRAM Mistral Small 3.2 24B needs

Memory Mistral Small 3.2 24B needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
16-bit44.7 GB46.8 GB50.6 GB65.6 GB
8-bit23.8 GB25.9 GB29.6 GB44.6 GB
6-bit18.3 GB20.4 GB24.2 GB39.2 GB
5-bit15.9 GB18.0 GB21.8 GB36.8 GB
4-bit13.6 GB15.7 GB19.4 GB34.4 GB
3-bit10.9 GB13.0 GB16.8 GB31.8 GB
2-bit8.4 GB10.5 GB14.2 GB29.2 GB
1.6-bit4.6 GB6.7 GB10.5 GB25.5 GB

One conversation, 16-bit context cache, runtime overhead included.

GPUs and Macs that run Mistral Small 3.2 24B

The best way Mistral Small 3.2 24B runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 4060Too bigPDF: Mistral Small 3.2 24B on RTX 4060 report
RTX 5060Too bigPDF: Mistral Small 3.2 24B on RTX 5060 report
RTX 3070Too bigPDF: Mistral Small 3.2 24B on RTX 3070 report
RTX 3080 10 GBToo bigPDF: Mistral Small 3.2 24B on RTX 3080 10 GB report
RTX 3060 12 GBToo bigPDF: Mistral Small 3.2 24B on RTX 3060 12 GB report
Arc B580Too bigPDF: Mistral Small 3.2 24B on Arc B580 report
RTX 4070Too bigPDF: Mistral Small 3.2 24B on RTX 4070 report
RTX 4070 SuperToo bigPDF: Mistral Small 3.2 24B on RTX 4070 Super report
RTX 5070Too bigPDF: Mistral Small 3.2 24B on RTX 5070 report
RTX 4060 Ti 16 GB3-bit20K17 tok/s2.3 sPDF: Mistral Small 3.2 24B on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB3-bit20K27 tok/s2.1 sPDF: Mistral Small 3.2 24B on RTX 5060 Ti 16 GB report
Arc A770 16 GB3-bit20K23 tok/s2.4 sPDF: Mistral Small 3.2 24B on Arc A770 16 GB report
Radeon RX 9070 XT3-bit20K32 tok/s1.4 sPDF: Mistral Small 3.2 24B on Radeon RX 9070 XT report
RTX 4070 Ti Super3-bit20K39 tok/s1.1 sPDF: Mistral Small 3.2 24B on RTX 4070 Ti Super report
RTX 4080 Super3-bit20K43 tok/s1.0 sPDF: Mistral Small 3.2 24B on RTX 4080 Super report
RTX 5070 Ti3-bit20K52 tok/s1.1 sPDF: Mistral Small 3.2 24B on RTX 5070 Ti report
RTX 50803-bit20K55 tok/s0.9 sPDF: Mistral Small 3.2 24B on RTX 5080 report
Radeon RX 7900 XTX6-bit20K28 tok/s2.2 sPDF: Mistral Small 3.2 24B on Radeon RX 7900 XTX report
RTX 30906-bit20K33 tok/s1.4 sPDF: Mistral Small 3.2 24B on RTX 3090 report
RTX 40906-bit20K35 tok/s0.6 sPDF: Mistral Small 3.2 24B on RTX 4090 report
Radeon AI PRO R97008-bit36K15 tok/s1.4 sPDF: Mistral Small 3.2 24B on Radeon AI PRO R9700 report
RTX 50908-bit36K48 tok/s0.5 sPDF: Mistral Small 3.2 24B on RTX 5090 report
2× RTX 30908-bit128K26 tok/s0.8 sPDF: Mistral Small 3.2 24B on 2× RTX 3090 report
2× RTX 40908-bit128K28 tok/s0.3 sPDF: Mistral Small 3.2 24B on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit128K36 tok/s0.7 sPDF: Mistral Small 3.2 24B on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit128K26 tok/s0.5 sPDF: Mistral Small 3.2 24B on RTX 6000 Ada report
L40S8-bit128K24 tok/s0.9 sPDF: Mistral Small 3.2 24B on L40S report
2× RTX 50908-bit128K48 tok/s0.3 sPDF: Mistral Small 3.2 24B on 2× RTX 5090 report
4× RTX 30908-bit128K26 tok/s0.4 sPDF: Mistral Small 3.2 24B on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit128K48 tok/s0.4 sPDF: Mistral Small 3.2 24B on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: Mistral Small 3.2 24B on Mac mini M4 16 GB report
Mac mini M4 24 GB3-bit20K8.4 tok/s32 sPDF: Mistral Small 3.2 24B on Mac mini M4 24 GB report
Mac mini M4 32 GB5-bit20K5.8 tok/s32 sPDF: Mistral Small 3.2 24B on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB3-bit20K18 tok/s16 sPDF: Mistral Small 3.2 24B on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit60K8.8 tok/s16 sPDF: Mistral Small 3.2 24B on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit128K8.8 tok/s16 sPDF: Mistral Small 3.2 24B on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit128K16 tok/s7.9 sPDF: Mistral Small 3.2 24B on M4 Max 128 GB report
M5 Max 128 GB8-bit128K18 tok/s2.3 sPDF: Mistral Small 3.2 24B on M5 Max 128 GB report
M3 Ultra 96 GB8-bit128K20 tok/s4.4 sPDF: Mistral Small 3.2 24B on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit128K20 tok/s4.4 sPDF: Mistral Small 3.2 24B on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit128K20 tok/s4.4 sPDF: Mistral Small 3.2 24B on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit128K29 tok/s1.4 sPDF: Mistral Small 3.2 24B on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit128K6.1 tok/s4.5 sPDF: Mistral Small 3.2 24B on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit128K6.1 tok/s4.5 sPDF: Mistral Small 3.2 24B on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit128K8.4 tok/s1.6 sPDF: Mistral Small 3.2 24B on RTX Spark report
DGX Spark8-bit128K7.6 tok/s1.6 sPDF: Mistral Small 3.2 24B on DGX Spark report
Dual DGX Spark8-bit128K7.6 tok/s0.9 sPDF: Mistral Small 3.2 24B on Dual DGX Spark report
DGX Station GB3008-bit128K162 tok/s0.1 sPDF: Mistral Small 3.2 24B on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit128K48 tok/s0.2 sPDF: Mistral Small 3.2 24B on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit128K48 tok/s0.1 sPDF: Mistral Small 3.2 24B on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit128K48 tok/s0.1 sPDF: Mistral Small 3.2 24B on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit128K54 tok/s1.0 sPDF: Mistral Small 3.2 24B on A100 80 GB SXM report
H100 SXM8-bit128K85 tok/s0.3 sPDF: Mistral Small 3.2 24B on H100 SXM report
4× H200 NVL8-bit128K117 tok/s0.1 sPDF: Mistral Small 3.2 24B on 4× H200 NVL report
8× H200 NVL8-bit128K117 tok/s0.1 sPDF: Mistral Small 3.2 24B on 8× H200 NVL report

Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.

How to run it

Ollama
ollama run mistral-small3.2:24b
llama.cpp
llama-server -hf unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF
vLLM
vllm serve mistralai/Mistral-Small-3.2-24B-Instruct-2506
MLX (Mac)
mlx_lm.chat --model mlx-community/Mistral-Small-3.2-24B-Instruct-2506-4bit

Questions

How much VRAM does Mistral Small 3.2 24B need?

Mistral Small 3.2 24B needs 13.6 GB for its weights at 4-bit, 15.7 GB with an 8K context and 19.4 GB with 32K. At 8-bit, 25.9 GB.

What is the minimum VRAM for Mistral Small 3.2 24B?

About 13.0 GB at 3-bit with an 8K context, the lowest precision we recommend for a model this size. The smallest machine here that runs it without spilling is the RTX 4060 Ti 16 GB.

How big is Mistral Small 3.2 24B?

24B parameters: 44.7 GB of weights at 16-bit, 13.6 GB at 4-bit.

Can Mistral Small 3.2 24B run on an RTX 4090?

Yes. At 6-bit it holds 20K of context and writes about 35 tokens a second.

More models