Loading…
Loading…
Models / Kimi-K2
2 models (~1.03e+03B to ~1.03e+03B) · ↓ 367.6K downloads a month · from $79.27/h on 8x Nvidia B200 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalmoonshotai/Kimi-K2-Instruct | vLLM | 1029 GB | 1161.7 GB | $79.27/h 8x Nvidia B200 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalmoonshotai/Kimi-K2-Base | vLLM | 1029 GB | 1161.7 GB | $79.27/h 8x Nvidia B200 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").