How much VRAM does Mistral Small 4 119B need?

Mistral Small 4 119B (119B parameters) requires approximately 84.0 GB of memory with Q4_K_M quantization.

What is the best quantization for Mistral Small 4 119B?

The recommended quantization for Mistral Small 4 119B is Q4_K_M, which balances quality and memory efficiency.

Can it run?

Yes, NVIDIA H100 80GB can run Mistral Small 4 119B with a C grade (Runs with offload (needs ~3.5 GB host RAM)). Expected decode speed: 110.1 tok/s.

CUsable

Runs with offload (needs ~3.5 GB host RAM)

Using Q4_K_M in vLLM

Capabilities:

Fit status

Runs with offload (needs ~3.5 GB host RAM)

Decode

110.1 tok/s

TTFT

1758 ms

Safe context

15K

Memory

84.0 GB / 80.0 GB

Weights72.6 GB

KV Cache1.0 GB

Runtime2.4 GB

Headroom8.0 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	C	Runs with offload (needs ~4.3 GB host RAM)	109.5 tok/s	2573 ms	30K
Chat	C	Runs with offload (needs ~3.3 GB host RAM)	110.2 tok/s	958 ms	8K
Coding	C	Runs with offload (needs ~3.5 GB host RAM)	110.1 tok/s	1758 ms	15K
RAG	C	Runs with offload (needs ~4.3 GB host RAM)	109.5 tok/s	3216 ms	30K
Reasoning	C	Runs with offload (needs ~3.5 GB host RAM)	110.1 tok/s

How Mistral Small 4 119B (119B params) fits at each quantization level on NVIDIA H100 80GB (80.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	46.4 GB	Low	C42
Q3_K_SBest for your GPU	3	58.3 GB	Low	C45

HuggingFace

huggingface-cli download mistral-small-4-119b

Upgrade options

AMD Instinct MI250X 128GBBudget pick

B100 tok/s decode

~$15,000 MSRP

NVIDIA H200 141GBBest value

B161.6 tok/s decode

~$30,000 MSRP

NVIDIA H200 PCIe 141GBNVIDIA upgrade

B161.6 tok/s decode

~$30,000 MSRP

Gaudi 3 128GB