axteam/Qwen3.5-9B-FP8-dynamic
Image-Text-to-Text • 10B • Updated • 29 • 3
Qwen
Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-9B-AWQ-8bit
Image-Text-to-Text • 9B • Updated • 27 • 3Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-9B-NVFP4A16
Image-Text-to-Text • 9B • Updated • 29 • 4Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-9B-AWQ-4bit
Image-Text-to-Text • 9B • Updated • 22 • 1Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-9B-NVFP4
Image-Text-to-Text • 9B • Updated • 23 • 1Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-9B-GGUF
9B • Updated • 126 • 2Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.8-27B-FP8-dynamic
Image-Text-to-Text • 28B • Updated • 33 • 1Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.8-27B-AWQ-8bit
Image-Text-to-Text • 28B • Updated • 28 • 1Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.8-27B-NVFP4A16
Image-Text-to-Text • 28B • Updated • 38 • 1Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.8-27B-AWQ-4bit
Image-Text-to-Text • 28B • Updated • 29 • 1Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.8-27B-NVFP4
Image-Text-to-Text • 28B • Updated • 43 • 1Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.8-27B-GGUF
27B • Updated • 120 • 2Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.6-35B-A3B-FP8-dynamic
Image-Text-to-Text • 36B • Updated • 29 • 1Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.6-35B-A3B-AWQ-8bit
Image-Text-to-Text • 36B • Updated • 19 • 1Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.6-35B-A3B-AWQ-4bit
Image-Text-to-Text • 36B • Updated • 24 • 1Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.6-35B-A3B-NVFP4
Image-Text-to-Text • 36B • Updated • 35 • 1Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.6-35B-A3B-GGUF
36B • Updated • 113 • 1Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.5-122B-A10B-AWQ-8bit
Image-Text-to-Text • 125B • Updated • 17 • 1Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-122B-A10B-AWQ-4bit
Image-Text-to-Text • 125B • Updated • 17 • 1Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-122B-A10B-FP8-dynamic
Image-Text-to-Text • 125B • Updated • 25 • 1Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-122B-A10B-NVFP4
Image-Text-to-Text • 125B • Updated • 27 • 1Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-122B-A10B-GGUF
125B • Updated • 123 • 1Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.6-35B-A3B-NVFP4A16
Image-Text-to-Text • 36B • Updated • 28 • 1Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-122B-A10B-NVFP4A16
125B • Updated • 23 • 2Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-0.8B-FP8-dynamic
Image-Text-to-Text • 0.9B • Updated • 18Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-0.8B-AWQ-8bit
Image-Text-to-Text • 0.9B • Updated • 19Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-0.8B-NVFP4A16
Image-Text-to-Text • 0.9B • Updated • 22Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-0.8B-AWQ-4bit
Image-Text-to-Text • 0.9B • Updated • 18Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-0.8B-NVFP4
Image-Text-to-Text • 0.9B • Updated • 18Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-0.8B-GGUF
0.8B • Updated • 112Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.5-2B-FP8-dynamic
Image-Text-to-Text • 2B • Updated • 18Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-2B-AWQ-8bit
Image-Text-to-Text • 2B • Updated • 19Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-2B-NVFP4A16
Image-Text-to-Text • 2B • Updated • 19Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-2B-AWQ-4bit
Image-Text-to-Text • 2B • Updated • 15Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-2B-NVFP4
Image-Text-to-Text • 2B • Updated • 15Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-2B-GGUF
2B • Updated • 107Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.5-4B-FP8-dynamic
Image-Text-to-Text • 5B • Updated • 14Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-4B-AWQ-8bit
Image-Text-to-Text • 5B • Updated • 19Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-4B-NVFP4A16
Image-Text-to-Text • 5B • Updated • 18Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-4B-AWQ-4bit
Image-Text-to-Text • 5B • Updated • 72Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-4B-NVFP4
Image-Text-to-Text • 5B • Updated • 21Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-4B-GGUF
4B • Updated • 105Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama
axteam/Qwen3.5-27B-FP8-dynamic
Image-Text-to-Text • 28B • Updated • 15Note FP8 dynamic (W8A8) — near-lossless 8-bit, vLLM
axteam/Qwen3.5-27B-AWQ-8bit
Image-Text-to-Text • 28B • Updated • 17Note AWQ INT8 (W8A16) — near-lossless 8-bit
axteam/Qwen3.5-27B-NVFP4A16
Image-Text-to-Text • 28B • Updated • 21Note NVFP4 W4A16 (FP4 weights, BF16 act) — best 4-bit quality
axteam/Qwen3.5-27B-AWQ-4bit
Image-Text-to-Text • 28B • Updated • 23Note AWQ INT4 (W4A16) — ~1/4 size
axteam/Qwen3.5-27B-NVFP4
Image-Text-to-Text • 28B • Updated • 36Note NVFP4 W4A4 (FP4 weights+act) — fastest on Blackwell
axteam/Qwen3.5-27B-GGUF
27B • Updated • 112Note GGUF Q8_0 / Q4_K_M for llama.cpp / Ollama