Loading…
Loading…
Models / KAT-Coder-V2.5-Dev
1 model (~34.7B) · ↓ 81.2K downloads a month · from $0.10/h on RTX 3060 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-Coder-V2.5-Dev | vLLM | 69 GB | 80 GB | $0.89/h 2x Nvidia A40 |
| 35B A3B ABLITERATED UNCENSORED PHILADELPHIA CLASS (KridgeDookie)KridgeDookie/KAT-Coder-V2.5-Dev-35B-A3B-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS | vLLM | 69 GB | 80 GB | $0.89/h 2x Nvidia A40 |
| MTP APEX GGUF (Myric)Myric/KAT-Coder-V2.5-Dev-MTP-APEX-GGUF | vLLM | 1.7 GB | 4.8 GB | $0.10/h RTX 3060 |
| REAP 50 NVFP4A16 (Ttimms)Ttimms/KAT-Coder-V2.5-Dev-REAP-50-NVFP4A16 | vLLM | 13 GB | 17.8 GB | $0.16/h 2x RTX 3060 |
| AWQ INT4 (cyankiwi)cyankiwi/KAT-Coder-V2.5-Dev-AWQ-INT4 | vLLM | 24 GB | 30.1 GB | $0.45/h 2x RTX 3090 |
| NVFP4A16 (sahilchachra)sahilchachra/KAT-Coder-V2.5-Dev-NVFP4A16 | vLLM | 22 GB | 27.9 GB | $0.31/h 2x RTX A4000 |
| W4A16 (sahilchachra)sahilchachra/KAT-Coder-V2.5-Dev-W4A16 | vLLM | 21 GB | 26.2 GB | $0.31/h 2x RTX A4000 |
| AutoRound W4A16 Tuning (LeaderboardModel1)LeaderboardModel1/KAT-Coder-V2.5-Dev-AutoRound-W4A16-Tuning | vLLM | 20 GB | 25.7 GB | $0.31/h 2x RTX A4000 |
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").