Mistral-Large-4-1T-A52B: hardware requirements
Mistral-Large-4-1T-A52B needs about 643 GB of memory at Q4_K_M quantization: a 635 GB download plus room for context and a runtime allowance. That is beyond consumer hardware; use it through an API instead.
This is a mixture-of-experts model. Only a fraction of its 1000B parameters are used for each token, so it runs faster than its size suggests, but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →
Will it run on your machine?
Use it without your own hardware
OpenRouter lists Mistral Large 4 at $0.70 per million input tokens and $2.10 per million output tokens. Artificial Analysis measured hosted providers generating 161 tokens per second (Oct 30, 2026).
Price as listed on OpenRouter today. Speed is for hosted servers, not a home computer.
Download sizes by quantization
Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?
| Quant | Download | RAM @ 4K | RAM @ 32K | Source |
|---|---|---|---|---|
| Q4_K_Mrecommended | 635 GB | 643 GB | 677 GB | ggml-org/Mistral-Large-4-1T-A52B-GGUF |
| Q8_0 | 1115 GB | 1123 GB | 1157 GB | ggml-org/Mistral-Large-4-1T-A52B-GGUF |
| BF16 | 2102 GB | 2110 GB | 2144 GB | ggml-org/Mistral-Large-4-1T-A52B-GGUF |
Download sizes are measured from the published files named in the Source column. The RAM columns add the context cache, calculated from the model's configuration, and a runtime allowance of 10% of the download (at least 0.5GB, at most 3GB).