88plug/Qwen3.6-35B-A3B-W4A16
Image-Text-to-Text • 35B • Updated • 594
88plug W4A16 Quantizations
INT4 weight-only (W4A16) compressed-tensors. Gold path: AutoRound iters=200 where stamped. Native vLLM.
-
-
88plug/Qwen3.6-27B-W4A16
Image-Text-to-Text • 27B • Updated • 507 -
88plug/Qwen3-Omni-30B-W4A16
Text Generation • 35B • Updated • 1.48k • 1 -
88plug/Qwen2.5-Omni-7B-W4A16
Text Generation • 11B • Updated • 3.58k -
88plug/Gemma4-E4B-it-W4A16
Image-Text-to-Text • 8B • Updated • 244 -
88plug/Gemma4-E2B-it-W4A16
Image-Text-to-Text • 5B • Updated • 652 -
88plug/MiniCPM-o-4.5-W4A16
Image-Text-to-Text • 9B • Updated • 604 • 1 -
88plug/Nemotron-3-Nano-30B-A3B-W4A16
Text Generation • 33B • Updated • 2.55k -
88plug/Kimi-VL-A3B-Thinking-2506-W4A16
Image-Text-to-Text • 17B • Updated • 830