WeiboAI/VibeThinker-3B
Full fine-tune vs Qwen/Qwen2.5-Coder-3B. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 73/73, MLP 108/108, Attention 252/252
33 derivatives checked
- prithivMLmods/VibeThinker-3B-GGUF: Same weights, converted
- squ11z1/Mythos-nano: Partial fine-tune
- mradermacher/VibeThinker-3B-i1-GGUF: Quantization
- constructai/VibeThinker-3B-GGUF: Same weights, converted
- marcosvpj/Mythos-nano: Partial fine-tune
- Shadow0482/mythos_fast: LoRA-style
- mradermacher/VibeThinker-3B-GGUF: Same weights, converted
- RefinedNeuro/VibeThinker-3B-Hermes-GGUF: LoRA-style
- Hermecente/Mythos-nano: Partial fine-tune
- DhruvalLabs/VibeThinker-3B-GGUF: Quantization
- KakTakOne/VibeThinker-3B-GGUF: Same weights, converted
- usermma/VibeThinker-3B-GGUF: Same weights, converted
- ermiaazarkhalili/VibeThinker-3B-SFT-Claude-Opus-Reasoning-Unsloth-GGUF: Full fine-tune
- ermiaazarkhalili/VibeThinker-3B-Function-Calling-xLAM-Unsloth-GGUF: Full fine-tune
- RefinedNeuro/RefinedToolCallV5-3b: LoRA-style
- richardyoung/Mythos-nano-heretic: Partial fine-tune
- mlx-community/VibeThinker-3B-OptiQ-4bit: Quantization
- Nemesispro/Mythos-nano: Partial fine-tune
- N14QM/Spark-X2.5-1.7B-VibeThinker-VN: Not comparable yet
- void0x14/Mythos-nano: Partial fine-tune
- oussaber/VibeThinker-3B-Q4_K_M-GGUF: Quantization
- orangejuicesmith/Mythos-Nano: Partial fine-tune
- mlx-community/VibeThinker-3B-8bit: Quantization
- heath0xFF/VibeThinker-3B-GGUF: Same weights, converted
- isyuricunha/Mythos-nano: Partial fine-tune
- mlx-community/VibeThinker-3B: Same weights, repacked
- mlx-community/VibeThinker-3B-4bit: Quantization
- ulamai/Ulam-1-Small: Full fine-tune
- prithivMLmods/VibeThinker-3B-heretic_decensored: Abliterated
- masterman331/VibeThinker-3B-Heretic-Strong: Abliterated
3 other repos hold byte-for-byte identical weight files.