Qwen/Qwen3.5-9B
Full fine-tune vs Qwen/Qwen3.5-9B-Base. Fine-tune: 96% of the weights changed, across 32 of 32 layers.
Tensors changed: lm_head 1/1, Embeddings 3/5, MLP 103/211, Attention 201/336, other 1/1, Norms 113/221
606 derivatives checked
- unsloth/Qwen3.5-9B-GGUF: Same weights, converted
- HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive: LoRA-style
- lmstudio-community/Qwen3.5-9B-MLX-4bit: Quantization
- lmstudio-community/Qwen3.5-9B-MLX-8bit: Quantization
- QuantTrio/Qwen3.5-9B-AWQ: Quantization
- lmstudio-community/Qwen3.5-9B-GGUF: Quantization
- AxionML/Qwen3.5-9B-NVFP4: Quantization
- autotrust/JEV-9B: Same weights, repacked (fewer layers)
- bartowski/Qwen_Qwen3.5-9B-GGUF: Same weights, converted
- unsloth/Qwen3.5-9B: Identical copy
- cyankiwi/Qwen3.5-9B-AWQ-4bit: Quantization
- unsloth/Qwen3.5-9B-MTP-GGUF: Same weights, converted
- ig1/Qwen3.5-9B-NVFP4: Quantization
- RedHatAI/Qwen3.5-9B-quantized.w8a8: Quantization
- RedHatAI/Qwen3.5-9B-FP8-dynamic: Quantization
- lukey03/Qwen3.5-9B-abliterated-GGUF: LoRA-style
- mlx-community/Qwen3.5-9B-MLX-4bit: Quantization
- INCModel3/Qwen3.5-9B-MXFP8-FP8KV-FP8Attn-CT-AutoRound: Quantization
- RedHatAI/Qwen3.5-9B-quantized.w4a16: Quantization
- XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B: Full fine-tune
- Cloudflare/clef-flash: LoRA-style
- Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2-GGUF: Full fine-tune
- Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-GGUF: Full fine-tune
- samuelchristlie/Ornith-1.0-9B-gguf: Full fine-tune
- inclusionAI/UI-Venus-2-9B: Full fine-tune
- lovedheart/Qwen3.5-9B-FP8: Quantization
- prithivMLmods/Gliese-Qwen3.5-9B-Abliterated-Caption: Abliterated
- Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled: LoRA-style
- mindchain/imajev-9b-GGUF: Full fine-tune
- empero-ai/Qwen3.8-9B-Distill: Partial fine-tune
92 other repos hold byte-for-byte identical weight files.