Loading…
Loading…
Models / KAT-V1
GPU memory needed (smallest version of each): KAT-V1 40B about 95.4 GB. The cheapest runs on RTX PRO 6000 from $1.53/h right now.
1 model (40B) · ↓ 164 downloads a month · from $1.53/h on RTX PRO 6000 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 |
|---|---|---|---|---|
| OriginalKwaipilot/KAT-V1-40B | vLLM | 81 GB | 95.4 GB | $1.53/h RTX PRO 6000 |
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").