Rent a GPU

Mistral Large 4 1T-A52B hardware requirements

1000B parameters (52B active, mixture of experts). Needs about 607.5 GB at Q4_K_M with 8K context.

Parameters1000B52B active (MoE)
Q4_K_M · 8K607.5 GBthe usual download
FP16 · 8K2001.3 GBfull precision
Runs well on0/51GPUs and Macs we track

Memory needed at each quantization

Weights, KV cache and runtime overhead, at 8K and 32K tokens of context.

QuantQuality8K context32K context
FP16Full quality2001.3 GB2003.3 GB
Q8_0Near-lossless1063.8 GB1065.8 GB
Q6_KExcellent821.3 GB823.3 GB
Q5_K_MVery good712.5 GB714.6 GB
Q4_K_MGood, most popular607.5 GB609.6 GB
Q3_K_MNoticeably degraded490 GB492.1 GB
Q2_KHeavily degraded376.3 GB378.3 GB

Model card

mistralai/Mistral-Large-4-1T-A52B on Hugging Face: licence, prompt format and GGUF builds.

Run it in the cloud

Run Mistral Large 4 1T-A52B on a rented GPU today.

It needs about 607.5 GB at Q4_K_M: a rented 8× A100 / H100 runs it for about $8.88 to $31.92 an hour. Compare with Claude and APIs or see how it scores against them.

RunPodRent a GPU on RunPodVast.aiRent a GPU on Vast.ai

Every GPU and Mac checked

Sorted by memory. Bars show how much of each card Mistral Large 4 1T-A52B fills at the best quant that fits.

51 Too big

NameMemory usedBest quantSpeed

Other Mistral models