Question 1

How much VRAM does Mistral-Large-4-1T-A52B need?

Accepted Answer

Mistral-Large-4-1T-A52B (52B) requires approximately 116.22905731201172 GB of VRAM at full precision (FP16), 58.1 GB at 8-bit quantization (Q8), and 29.1 GB at 4-bit quantization (Q4). The actual VRAM usage may be slightly higher due to KV cache and runtime overhead.

Question 2

What GPUs can run Mistral-Large-4-1T-A52B?

Accepted Answer

We have 9 compatible GPUs indexed for Mistral-Large-4-1T-A52B, including H200 SXM, MI300X, MI350X, B200, B100 and 4 more. Compatibility is determined by whether the GPU's VRAM can fit the model's memory footprint, potentially using quantization to reduce memory usage.

Question 3

What is the cheapest GPU to run Mistral-Large-4-1T-A52B?

Accepted Answer

The cheapest GPU to run Mistral-Large-4-1T-A52B in the cloud is the H200 SXM at $2.35/hr. Other affordable options include MI300X ($2.39/hr) and MI350X ($5.49/hr). Prices reflect the lowest on-demand rate across cloud providers.

Question 4

Can Mistral-Large-4-1T-A52B run on consumer hardware?

Accepted Answer

Mistral-Large-4-1T-A52B requires approximately 116.22905731201172 GB of VRAM at FP16, and even with 4-bit quantization (Q4) it needs about 29.1 GB. This exceeds the 24 GB available on the highest-end consumer GPU (RTX 4090), so Mistral-Large-4-1T-A52B requires datacenter or cloud GPUs.

Question 5

What quantization options are available for Mistral-Large-4-1T-A52B?

Accepted Answer

Mistral-Large-4-1T-A52B supports the following quantization levels: FP16 (full precision): ~116.22905731201172 GB, Q8 (8-bit): ~58.1 GB, Q6 (6-bit): ~43.6 GB, Q4 (4-bit): ~29.1 GB. Lower quantization reduces VRAM usage at the cost of some quality loss. Q8 offers a good balance between quality and memory savings, while Q4 provides maximum compression.

GPU	VRAM	Best Quant	From	Rent
H200 SXM	141 GB	FP16	$2.35/hr	Rent
MI300X	192 GB	FP16	$2.39/hr	Rent
MI350X	288 GB	FP16	$5.49/hr	Rent
B200	192 GB	FP16	$6.99/hr
B100	192 GB	FP16	—
MI325X	256 GB	FP16	—
Gaudi 3	128 GB	Q8	—
DGX Spark (GB10)	128 GB	Q8	—
M3 Ultra (192 GB)	144 GB	FP16	—

Mistral-Large-4-1T-A52B

VRAM Requirements

Compatible GPUs

Mistral-Large-4-1T-A52B FAQ

VRAM requirements for Mistral-Large-4-1T-A52B at different quantization levels
Quantization	VRAM Required
FP16	116.2 GB
Q8	58.1 GB
Q6	43.6 GB
Q4	29.1 GB