This page answers AI Mo Kimina Prover 72B q2_k quantization queries with explicit calculations from our model requirement dataset and compatibility speed table.
Short answer: AI Mo Kimina Prover 72B typically needs around 20GB VRAM at Q2_K, and 24GB is safer for smoother usage.
Estimated from Q4 using a 45% memory reduction assumption for Q2_K.
Throughput data below uses available compatibility measurements/estimates and is sorted by tokens per second for this model.
Need general guidance? Review full methodology.
| GPU | VRAM | Quantization | Speed | Compatibility | Buy |
|---|---|---|---|---|---|
| AMD Instinct MI300X | 192GB | Q4 | 153 tok/s | View full compatibility | Buy options |
| NVIDIA H200 SXM 141GB | 141GB | Q4 | 138 tok/s | View full compatibility | Buy options |
| NVIDIA H100 SXM5 80GB | 80GB | Q4 | 99 tok/s | View full compatibility | Buy options |
| AMD Instinct MI250X | 128GB | Q4 | 96 tok/s | View full compatibility | Buy options |
| NVIDIA H100 PCIe 80GB | 80GB | Q4 | 63 tok/s | View full compatibility | Buy options |
| RTX 5090 | 32GB | Q4 | 60 tok/s | View full compatibility | Buy options |
| NVIDIA A100 80GB SXM4 | 80GB | Q4 | 58 tok/s | View full compatibility | Buy options |
| AMD Instinct MI210 | 64GB | Q4 | 48 tok/s | View full compatibility | Buy options |
| NVIDIA A100 40GB PCIe | 40GB | Q4 | 45 tok/s | View full compatibility | Buy options |
| RTX 4090 | 24GB | Q4 | 36 tok/s | View full compatibility | Buy options |
| NVIDIA RTX 6000 Ada | 48GB | Q4 | 36 tok/s | View full compatibility | Buy options |
| NVIDIA L40 | 48GB | Q4 | 33 tok/s | View full compatibility | Buy options |
AI Mo Kimina Prover 72B at Q2_K is estimated to require about 20GB VRAM minimum, with 24GB recommended for smoother operation.
Start with AMD Instinct MI300X, NVIDIA H200 SXM 141GB, NVIDIA H100 SXM5 80GB and review each compatibility page for full speed and fit details.
Q2_K is a balance point between memory usage and quality. If your GPU is below 20GB, consider lower-bit quantization; if you have extra VRAM, compare Q8/FP16 options for quality-sensitive workloads.