AMD Instinct MI300X meets the minimum VRAM requirement for Q4 inference of moonshotai/Kimi-K2.5. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
AMD Instinct MI300X can run moonshotai/Kimi-K2.5 with Q4 quantization. At approximately 840 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 188GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 4GB | 192GB | 839.51 tok/s | ✅ Fits comfortably |
| Q8 | 8GB | 192GB | 493.70 tok/s | ✅ Fits comfortably |
| FP16 | 16GB | 192GB | 279.80 tok/s | ✅ Fits comfortably |