指南 / 模型家族

模型家族

同一基座衍生的微调 / LoRA / 量化模型放在一起看,共 45 个家族。灰色名字是站上未收录的基座。

20 个成员

Qwen3.8-27B
- 量化 Dirk-Qwen3.8-27B (peculiar-ragdoll)
- 量化 Huihui-Qwen3.8-27B-abliterated-GGUF (huihui-ai)
- 量化 Qwen3.8-27B (ISTA-DASLab)
- 量化 Qwen3.8-27B (LM Studio)
- 量化 Qwen3.8-27B (Qwen)
- 量化 Qwen3.8-27B Heretic Abliterated GGUF (0bserverx)
- 量化 Qwen3.8-27B MTPLX (Youssofal)
- LoRA Qwen3.8-27B-Atlassian-Q8-mlx (LeanZero)
- 微调 Qwen3.8-27B-DFlash2 (z-lab)
- 量化 Qwen3.8-27B-Escha-W2 (Escha)
- 量化 Qwen3.8-27B-MTPLX-Optimized-Speed (Youssofal)
- 量化 Qwen3.8-27B-OBLITERATED (OBLITERATUS)
- 量化 Qwen3.8-27B-Ridge-GGUF (empero-ai)
- 量化 Qwen3.8-27B-Uncensored-GGUF (orcarouter)
- 量化 Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF (HauhauCS)
- 量化 Qwen3.8-27B-VQ (TheDrainFlorist)
- 量化 Qwen3.8-27B-VQ-4.5bpw (Qwen)
- 量化 Qwen3.8-27B-VQ-4.8bpw (Qwen)
- 量化 Qwen3.8-Whittle-MoE-27B-A17.8B (logic65)
- 量化 Qwopus3.8-27B-Flash-GGUF (Jackrong)

18 个成员

Qwen3.6-35B-A3B
- 微调 Macaron-V1-Tall (MindLab)
- 量化 Qwen3.6-35B-A3B (AutomatosX)
- 量化 Qwen3.6-35B-A3B (Qwen)
- 量化 Qwen3.6-35B-A3B (andreaborio)
- 量化 Qwen3.6-35B-A3B-Escha-W2 (EschaLabs)
- 量化 Qwen3.6-35B-A3B-GGUF (Unsloth)
- 量化 Qwen3.6-35B-A3B-NVFP4 (NVIDIA)
- 量化 Qwen3.6-35B-A3B-OptiQ-4bit (mlx-community)
- 量化 Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
  - 量化 Qwen3.6-35B-A3B (LuffyTheFox)
  - 量化 Qwen3.6-35B-A3B (Qwen)
  - 量化 Qwen3.6-35B-A3B-Genesis-Hermes-V5-GGUF (LuffyTheFox)
- 量化 Qwen3.6-35B-A3B-VQ (aquaman164)
- 量化 Qwen3.6-35B-A3B-VQ-3.4bpw (TheDrainFlorist)
- 量化 Qwen3.6-35B-A3B-VQ-3.8bpw (TheDrainFlorist)
- 量化 Qwen3.6-35B-A3B-VQ-4.6bpw (TheDrainFlorist)
- 量化 Qwen3.6-35B-A3B-VQ-5.4bpw (TheDrainFlorist)
- 微调 XYZ-Aquila-mini (XYZAILab)

12 个成员

Qwen3.6-27B
- LoRA BTL-3 (Bad Theory Labs)
- 量化 Bonsai-27B-mlx-1bit (Prism ML)
- 量化 Qwen3.6-27B (OBLITERATUS)
- 量化 Qwen3.6-27B-GGUF (unsloth)
- 量化 Qwen3.6-27B-MTP-pi-tune (bytkim)
- 量化 Qwen3.6-27B-OptiQ-4bit (mlx-community)
- 量化 Qwen3.6-27B-Uncensored-Aggressive (HauhauCS)
- LoRA Qwen3.6-27b-Fable5 (hotdogs)
- 量化 Qwopus3.6-27B-Fusion (KyleHessling1)
- LoRA Qwopus3.6-27B-v2-MTP-GGUF (Jackrong)
- 量化 Ternary-Bonsai-27B (Prism ML)
- 微调 grug-27b (ProCreations)

10 个成员

Qwen3.5-9B
- 量化 MaralGPT-Mythos-9B (MaralGPT)
- 量化 Qwen3.5-9B (Qwen)
- LoRA Qwen3.5-9B-Atlassian-Q4-mlx (LeanZero)
- 量化 Qwen3.5-9B-MTP-GGUF (unsloth)
- 量化 Qwen3.5-9B-OptiQ-4bit (mlx-community)
- 微调 Qwen3.8-9B (Empero)
- 微调 Qwen3.8-9B-Distill (Empero)
- 微调 Qwythos-9B (Empero)
  - 微调 Qwythos-9B-v2 (Empero AI)
- 微调 WeMM-Embedding-9B (Tencent)

6 个成员

MiniMax-H3
- 微调 FastVideo-FastH3-4-step-Preview-v1-VSA-DataFree (FastVideo)
- LoRA MiniMax H3 Realism People LoRA (fal)
- 量化 MiniMax-H3 (DeepBeepMeep)
- LoRA MiniMax-H3-RAVEN-Streaming-LoRA (mvp-lab)
- 微调 MiniMax-H3-x-Z-Image-native (MiniMaxAI)
- 微调 VDN-Minimax-H3 (OpenVDN)

6 个成员

Ornith-1.0-35B (DeepReinforce)
- 量化 Ornith-1.0-35B (deepreinforce-ai)
- 量化 Ornith-1.0-35B-MLX (LM Studio)
- 量化 Ornith-1.0-35B-MLX (deepreinforce-ai)
- 量化 Ornith-35B-MLX (AtomicChat)
- 量化 Ornith-35b-MLX-6bit (AtomicChat)

6 个成员

gemma-4-12B-it
- 微调 Gemma-4-12B Agentic v2 (yuxinlu1)
- 量化 Gemma-4-12B-OBLITERATED (OBLITERATUS)
- 量化 Gemma4-12B-Coder (yuxinlu1)
- 量化 Gemma4-12B-Uncensored (HauhauCS)
- 微调 Grug-12B (kai-os)
- 量化 gemma-4-12B-it (Google)

4 个成员

LFM2.5-2.6B (LiquidAI)
- 量化 LFM2.5-2.6B (LiquidAI)
- 量化 LFM2.5-2.6B-MLX-5bit (LM Studio)
- 微调 fuse-1-Lite (Akahsizrr)

4 个成员

MiniMax-H3
- 微调 ClipProj-MiniMax-H3 (NicoLab28)
- 量化 MiniMax-H3 (MiniMax)
- LoRA MiniMax-H3 Turbo LoRA (drbaph)
- LoRA MiniMax-H3 Turbo LoRA (larryvrh)

4 个成员

gemma-4-26B-A4B-it
- 量化 Gemma4-26B-A4B-Uncensored-Balanced (HauhauCS)
- 量化 SuperGemma4-26B (Jiunsong)
- 量化 SuperGemma4-26B-Uncensored (Jiunsong)
- 量化 gemma-4-26B-A4B-it (Google)

3 个成员

GigaAM-Multilingual (ai-sage)
- 量化 GigaAM-Multilingual MLX INT8 (ai-babai)
- 量化 GigaAM-Multilingual-MLX INT4 (ai-babai)

3 个成员

Krea-2-Turbo (Krea)
- LoRA Krea2 Turbo Style Reference LoRA (Krea)
- LoRA M87 (mgwr)

3 个成员

LTX-2.3
- 微调 Dramabox (ResembleAI)
- LoRA LTX-Best-Face-ID (Alissonerdx)
- 量化 Sulphur-2-base (SulphurAI)

3 个成员

Ornith-1.0-9B (DeepReinforce)
- 量化 Ornith-1.0-9B-MLX (deepreinforce-ai)
- 量化 Ornith-1.0-9B-MLX-6bit (deepreinforce-ai)

3 个成员

Qwen2.5-Coder-3B
- 微调 VibeThinker-3B (WeiboAI)
  - 量化 Mythos-nano (squ11z1)
    - 量化 Mythos-nano (squ11z1)

3 个成员

Qwen3-0.6B
- 微调 OmniVoice (k2-fsa)
- 微调 S1-mini (Superwhisper)
- 量化 hamo-score-0.6b (HamoAI)

3 个成员

Qwen3.8-Flash-Next
- 量化 Qwen3.8-Flash-Next (Qwen)
- 量化 Qwen3.8-Flash-Next Optimized Speed (MTPLX)
- 量化 Qwen3.8-Flash-Next-Uncensored (orcarouter)

3 个成员

diffusiongemma-26B-A4B-it
- 量化 DiffusionGemma-26B-A4B-it (Google)
- 量化 DiffusionGemma-26B-A4B-it-NVFP4 (NVIDIA)
- 微调 OUI-1 (Thesys)

3 个成员

gemma-4-31B-it
- 微调 Gemma-4-31B-StyleTune (Gryphe)
- 量化 Gemma-4-31B-it (Google)
- 量化 Gemma-4-31B-it-GGUF (LM Studio)

2 个成员

DeepSeek-V4-Flash
- 量化 DeepSeek-V4-Flash GGUF (DeepSeek)

2 个成员

DeepSeek-V4-Flash-0731
- 量化 DeepSeek-V4-Flash-0731 (DeepSeek)
- 量化 DeepSeek-V4-Flash-Ablit (huihui-ai)

2 个成员

Granite-4.2-30B (IBM)
- 量化 Granite-4.2-30B-MLX-6bit (LM Studio)

2 个成员

Krea-2-Raw (Krea)
- LoRA Krea-2 Depth ControlNet (Patil)

2 个成员

LFM2.5-350M-Base
- 微调 LFM2.5-ColBERT-350M (LiquidAI)
- 微调 LFM2.5-Embedding-350M (LiquidAI)

2 个成员

LFM2.5-8B-A1B (LiquidAI)
- 量化 LFM2.5-8B-A1B-APEX (LiquidAI / LocalAI)

2 个成员

LTX2.3-10Eros
- 量化 LTX-2.5-uncensored-v1.1-FP8 (ChrisColeTech)
- 量化 LTX2.3-10Eros-GGUF (vantagewithai)

2 个成员

Laguna-S-2.1
- 量化 Laguna-S-2.1 (poolside)
- 量化 Laguna-S-2.1-NVFP4 (poolside)

2 个成员

MiniCPM5-1B (OpenBMB)
- 微调 MiniCPM5-1B-Thinking (OpenBMB)

2 个成员

Ministral-3-3B-Instruct-2512
- 微调 Nemotron-3-Embed-1B (NVIDIA)
  - 量化 Nemotron-3-Embed-1B-NVFP4 (NVIDIA)

2 个成员

NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
- LoRA Koleslaw-Nemotron-30B (ThunkAboutIt)
- 微调 PhoneLLM Alpha 1 (Pipecat)

2 个成员

Ornith-1.5-35B-A3B (Ornith)
- 量化 Tiel-Coder-35B-A3B-GGUF (peculiar-ragdoll)

2 个成员

Qwen3-4B-Instruct-2507
- 微调 FastContext-1.0-4B-RL (Microsoft)
- 微调 FastContext-1.0-4B-SFT (Microsoft)

2 个成员

Qwen3.5-4B
- 微调 NeoHorse-1-4B (TokenRhythm)
- 微调 NuExtract3 (NuMind)

2 个成员

Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16
- 量化 Qwen3.6-27B-AEON-Uncensored (AEON-7)
- 量化 Qwen3.6-27B-Uncensored (AEON-7/mradermacher)

2 个成员

Qwen3.6-27B-Heretic2-Uncensored-Finetune-Thinking
- 量化 BugTraceAI-CORE-Ultra-27B (BugTraceAI)
- 量化 Qwen3.6-27B-Heretic-Uncensored-FINETUNE-NEO-CODE-Di-IMatrix-MAX-GGUF (DavidAU)

2 个成员

ThinkingCap-Qwen3.6-27B
- 量化 Dagger-Qwen3.6-27B (peculiar-ragdoll)
- 量化 ThinkingCap-Qwen3.6-27B (bottlecapai)

2 个成员

Z-Image
- 量化 Juggernaut-Z-Image (RunDiffusion)
- 量化 Z-Anime (SeeSee21)

2 个成员

bonsai-image-ternary-4B-unpacked
- 微调 Bonsai Image Ternary 4B (Prism ML)
- 量化 Bonsai-Image-4B (Prism ML)

2 个成员

gemma-4-E2B-it
- 量化 Gemma-4-12B GGUF 合集 (rahul7star)
- 量化 gemma-4-E2B-it-GGUF (LM Studio)

2 个成员

gemma-4-e4b-it
- 量化 Gemma-4-E4B-Aggressive (HauhauCS)
- 量化 gemma-4-E4B-it (Google)

2 个成员

gpt-oss-20b
- 量化 GPT-OSS-20B (OpenAI)
- 微调 Harness-1 (pat-jj)

2 个成员

granite-4.2-3b
- 量化 Granite-4.2-3B-MLX-6bit (IBM)
- 量化 Granite-4.2-3B-MLX-8bit (LM Studio)

2 个成员

granite-4.2-8b
- 量化 Granite-4.2-8B (IBM)
- 量化 Granite-4.2-8B-MLX-6bit (LM Studio)

2 个成员

ideogram-4-fp8 (ideogram-ai)
- LoRA ideogram_4_turbotime_lora (ostris)

2 个成员

prism-coder-14b
- 量化 prism-coder-14b (dcostenco)
- 量化 prism-coder-14b (mradermacher)