How much VRAM does Devstral Small 1.1 need?

Devstral Small 1.1 (24B parameters) requires approximately 26.0 GB of memory with Q4_K_M quantization.

What is the best quantization for Devstral Small 1.1?

The recommended quantization for Devstral Small 1.1 is Q4_K_M, which balances quality and memory efficiency.

Can it run?

Can NVIDIA A16 64GB run Devstral Small 1.1?

Q: Can NVIDIA A16 64GB run Devstral Small 1.1?

Yes, NVIDIA A16 64GB can run Devstral Small 1.1 with a C grade (Runs well). Expected decode speed: 29.5 tok/s.

CUsable

Runs well

Using Q4_K_M in Ollama

Capabilities:

Fit status

Runs well

Decode

29.5 tok/s

TTFT

6561 ms

Safe context

39K

Memory

26.0 GB / 64.0 GB

Memory breakdown

Weights14.6 GB

KV Cache3.8 GB

Runtime1.2 GB

Headroom6.4 GB

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Runs well	29.5 tok/s	9543 ms	69K
Chat	C	Runs well	29.5 tok/s	3579 ms	21K
Coding	C	Runs well	29.5 tok/s	6561 ms	39K
RAG	C	Runs well	29.5 tok/s	11929 ms	69K
Reasoning	C	Runs well	29.5 tok/s	7754 ms	39K

Quantization options

How Devstral Small 1.1 (24B params) fits at each quantization level on NVIDIA A16 64GB (64.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	9.4 GB	Low	D33
Q3_K_S	3	11.8 GB	Low	D33
NVFP4	4

Get started

Ollama

ollama run devstral-small-2507

HuggingFace

huggingface-cli download devstral-small-2507

Upgrade options

Hardware that runs Devstral Small 1.1 well

NVIDIA A100 80GBBudget pick

C117 tok/s decode

~$15,000 MSRP

AMD Instinct MI250X 128GBBest value

C170.5 tok/s decode

~$15,000 MSRP

NVIDIA H100 80GBNVIDIA upgrade

C192.2 tok/s decode

~$40,000 MSRP

NVIDIA A800 80GBBiggest leap

See all results for NVIDIA A16 64GB See all hardware for Devstral Small 1.1