NVIDIA H100 PCIe 80GB meets the minimum VRAM requirement for Q4 inference of AI-MO/Kimina-Prover-72B. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
NVIDIA H100 PCIe 80GB can run AI-MO/Kimina-Prover-72B with Q4 quantization. At approximately 66 tokens/second, you can expect Good speed - acceptable for interactive use.
You have 45GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 35GB | 80GB | 66.00 tok/s | ✅ Fits comfortably |
| Q8 | 70GB | 80GB | 40.14 tok/s | ✅ Fits comfortably |
| FP16 | 141GB | 80GB | 21.97 tok/s | ❌ Not recommended |