Mistral Large 4 (1T-A52B) VRAM & GPU Requirement Calculator
Interactive hardware sizing for Mistral Large 4 (1T-A52B) (1050B by Mistral AI). Estimated 8K context footprint: 531.0 GB (INT4) • 1056.7 GB (FP8) • 2108.2 GB (FP16).
Memory Requirements 1056.74 GB
Requires 14 GPUs (based on memory capacity)
1050 GB
All model weights
1.5 GB
Conversation history cache
4.37 GB
Expert model optimization
0.87 GB
Temporary computation cache
Estimated using memory bandwidth and tensor compute throughput across 14x NVIDIA H100 (80GB).
vllm serve mistralai/Mistral-Large-4-1T-A52B --quantization fp8 --tensor-parallel-size 14 --max-model-len 8192
Scenario Examples (GPU + Model + Concurrency):
Click these examples to quickly configure popular model deployment scenarios!
📋 Calculation Formula FAQ
How is GPU throughput calculated?
Why do we consider memory bandwidth limitations?
How are MoE models handled differently?
What does the system efficiency factor represent?
How does L2 cache hit rate affect GPU throughput calculations?
How does precision affect performance?
Why might my results differ from other calculators?
What are the budget-friendly GPU options for LLM deployment?
What are the enterprise-grade GPU solutions?
What are the GPU requirements for popular AI models?
What are the pro tips for GPU optimization?
Mistral Large 4 (1T-A52B) Architecture & Memory Profile
Mistral Large 4 (1T-A52B) VRAM Matrix by Precision & Context Length
Precomputed single-user memory requirements (model weights + KV cache + runtime overhead).
| Precision | 8K Context | 32K Context | 128K Context |
|---|---|---|---|
| FP16 | 2108.2 GB | 2119.3 GB | 2157.4 GB |
| FP8 | 1056.7 GB | 1063.3 GB | 1083.4 GB |
| INT8 | 1056.7 GB | 1063.3 GB | 1083.4 GB |
| INT4 | 531.0 GB | 535.3 GB | 546.4 GB |
Recommended GPU Hardware for Mistral Large 4 (1T-A52B) (FP8)
Minimum GPU count for 8K context at batch size 1 and estimated generation throughput.
| GPU Hardware | Min GPUs | Total VRAM Pool | Est. Speed |
|---|---|---|---|
| NVIDIA RTX 4090 (24GB) | 45× | 1080 GB | 23356 tok/s |
| NVIDIA RTX 5090 (32GB) | 34× | 1088 GB | 25030 tok/s |
| Apple M4 Max (128GB) (128GB) | 9× | 1152 GB | 434 tok/s |
| NVIDIA A100 (80GB) (80GB) | 14× | 1120 GB | 4972 tok/s |
| NVIDIA H100 (80GB) | 14× | 1120 GB | 8458 tok/s |
| NVIDIA H200 (141GB) | 8× | 1128 GB | 4268 tok/s |
| Huawei Ascend 910B (64GB) | 17× | 1088 GB | 4252 tok/s |