Uncensored CRACK builds of MiniCPM5 series (OpenBMB). Small dense text models with binary thinking mode.
dealign.ai PRO
dealignai
AI & ML interests
No cheap template tricks, no fine tuning here. Do the hard work of actually understanding.
Recent Activity
updated a model about 9 hours ago
dealignai/Xing4.0-29B-A4B-CRACK-GGUF updated a model 2 days ago
dealignai/GLM-5.3-Flash-CYBERSECURITY-W4A16 updated a model 2 days ago
dealignai/GLM-5.3-Flash-ABLITERATED-FP8Organizations
Bonsai 27B VLM CRACK models for Apple Silicon: one-bit and ternary JANG affine bundles with matched MMLU and full HB-320 results.
Abliterated Qwen 3.6 27B + 35B-A3B (MXFP4/8) with native d3 MTP, vision + video.
These are the empirically proven highest quality uncensored models on MLX.
-
dealignai/MiniMax-M2.7-JANGTQ-CRACK
Text Generation • 18B • Updated • 521 • 27 -
dealignai/Nemotron-3-Super-120B-A12B-UNCENSORED-JANG_2L
Text Generation • 51B • Updated • 2.41k • 14 -
dealignai/Gemma-4-26B-A4B-JANG_2L-CRACK
Image-Text-to-Text • 14B • Updated • 908 • 89 -
dealignai/Qwen3.5-VL-397B-A17B-UNCENSORED-JANG_1L
Image-Text-to-Text • 429B • Updated • 855 • 5
CRACK-abliterated Muse Glimmer 30B — vision, reasoning, tools. JANG MLX. HarmBench ~99.5%.
-
dealignai/Muse-Glimmer-30B-JANG_6M-CRACK
Image-Text-to-Text • 30B • Updated • 266 • 1 -
dealignai/Muse-Glimmer-30B-JANG_4M-CRACK
Image-Text-to-Text • 30B • Updated • 141 -
dealignai/Muse-Glimmer-30B-JANG_2D-CRACK
Image-Text-to-Text • 30B • Updated • 111 -
dealignai/Muse-Glimmer-30B-CRACK-GGUF
Image-Text-to-Text • 28B • Updated • 1.5k • 8
CRACK-abliterated Gemma 4 QAT — E2B/E4B/12B/26B-A4B/31B in MXFP4 & JANG_4M MLX quants. Vision, audio & reasoning intact. For vMLX.
-
dealignai/Gemma-4-E2B-it-qat-MXFP4-CRACK
Image-Text-to-Text • 5B • Updated • 111 • 2 -
dealignai/Gemma-4-E2B-it-qat-JANG_4M-CRACK
Image-Text-to-Text • 5B • Updated • 125 -
dealignai/Gemma-4-E4B-it-qat-MXFP4-CRACK
Image-Text-to-Text • 8B • Updated • 69 -
dealignai/Gemma-4-E4B-it-qat-JANG_4M-CRACK
Image-Text-to-Text • 8B • Updated • 214 • 2
CRACK-abliterated Qwen 3.5 and 3.6 models for MLX on Apple Silicon. Mixed-precision JANGTQ quantization with HarmBench 90%+ and MMLU-neutral profiles.
-
dealignai/Qwen3.6-35B-A3B-JANGTQ4-CRACK
Image-Text-to-Text • 9B • Updated • 391 • 12 -
dealignai/Qwen3.6-35B-A3B-JANGTQ2-CRACK
Image-Text-to-Text • 12B • Updated • 277 • 10 -
dealignai/Qwen3.6-27B-MXFP4-CRACK
Image-Text-to-Text • 27B • Updated • 159 • 4 -
dealignai/Qwen3.6-27B-JANG_4M-CRACK
Image-Text-to-Text • 27B • Updated • 718 • 24
CRACK abliterated MiniMax M2.5 REAP models — 139B and 172B at 4/6/8-bit MLX.
Uncensored CRACK builds of MiniCPM5 series (OpenBMB). Small dense text models with binary thinking mode.
CRACK-abliterated Muse Glimmer 30B — vision, reasoning, tools. JANG MLX. HarmBench ~99.5%.
-
dealignai/Muse-Glimmer-30B-JANG_6M-CRACK
Image-Text-to-Text • 30B • Updated • 266 • 1 -
dealignai/Muse-Glimmer-30B-JANG_4M-CRACK
Image-Text-to-Text • 30B • Updated • 141 -
dealignai/Muse-Glimmer-30B-JANG_2D-CRACK
Image-Text-to-Text • 30B • Updated • 111 -
dealignai/Muse-Glimmer-30B-CRACK-GGUF
Image-Text-to-Text • 28B • Updated • 1.5k • 8
Bonsai 27B VLM CRACK models for Apple Silicon: one-bit and ternary JANG affine bundles with matched MMLU and full HB-320 results.
CRACK-abliterated Gemma 4 QAT — E2B/E4B/12B/26B-A4B/31B in MXFP4 & JANG_4M MLX quants. Vision, audio & reasoning intact. For vMLX.
-
dealignai/Gemma-4-E2B-it-qat-MXFP4-CRACK
Image-Text-to-Text • 5B • Updated • 111 • 2 -
dealignai/Gemma-4-E2B-it-qat-JANG_4M-CRACK
Image-Text-to-Text • 5B • Updated • 125 -
dealignai/Gemma-4-E4B-it-qat-MXFP4-CRACK
Image-Text-to-Text • 8B • Updated • 69 -
dealignai/Gemma-4-E4B-it-qat-JANG_4M-CRACK
Image-Text-to-Text • 8B • Updated • 214 • 2
Abliterated Qwen 3.6 27B + 35B-A3B (MXFP4/8) with native d3 MTP, vision + video.
CRACK-abliterated Qwen 3.5 and 3.6 models for MLX on Apple Silicon. Mixed-precision JANGTQ quantization with HarmBench 90%+ and MMLU-neutral profiles.
-
dealignai/Qwen3.6-35B-A3B-JANGTQ4-CRACK
Image-Text-to-Text • 9B • Updated • 391 • 12 -
dealignai/Qwen3.6-35B-A3B-JANGTQ2-CRACK
Image-Text-to-Text • 12B • Updated • 277 • 10 -
dealignai/Qwen3.6-27B-MXFP4-CRACK
Image-Text-to-Text • 27B • Updated • 159 • 4 -
dealignai/Qwen3.6-27B-JANG_4M-CRACK
Image-Text-to-Text • 27B • Updated • 718 • 24
These are the empirically proven highest quality uncensored models on MLX.
-
dealignai/MiniMax-M2.7-JANGTQ-CRACK
Text Generation • 18B • Updated • 521 • 27 -
dealignai/Nemotron-3-Super-120B-A12B-UNCENSORED-JANG_2L
Text Generation • 51B • Updated • 2.41k • 14 -
dealignai/Gemma-4-26B-A4B-JANG_2L-CRACK
Image-Text-to-Text • 14B • Updated • 908 • 89 -
dealignai/Qwen3.5-VL-397B-A17B-UNCENSORED-JANG_1L
Image-Text-to-Text • 429B • Updated • 855 • 5
CRACK abliterated MiniMax M2.5 REAP models — 139B and 172B at 4/6/8-bit MLX.