Loading…
Loading…
Models / Keye-VL-1_5
GPU memory needed (smallest version of each): Keye-VL-1_5 8B about 24.2 GB. The cheapest runs on 2x RTX 5060 Ti from $0.39/h right now.
1 model (8B) · ↓ 13K downloads a month · from $0.39/h on 2x RTX 5060 Ti 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 |
|---|---|---|---|---|
| OriginalKwai-Keye/Keye-VL-1_5-8B | vLLM | 17 GB | 24.2 GB | $0.39/h 2x RTX 5060 Ti |
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").