Muse Glimmer NVFP4 W4A16 development conversion. Vision stays at source precision. Not quality-certified.
AutomatosX PRO
AutomatosX
AI & ML interests
None yet
Recent Activity
updated a collection 1 day ago
AXDocLayout updated a collection 2 days ago
Qwen updated a collection 2 days ago
Muse GlimmerOrganizations
Original AXDocLayout source and reference NVFP4/MXFP4/MXFP8 storage. Reference inference reconstructs FP32 PyTorch CPU weights.
MLX AXQ MXFP4 development checkpoints; reference-only storage is listed separately.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k
AXQuant CUDA NVFP4 W4A16/W4A4 development containers. Runtime evidence is pack-specific.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
Cloudflare Clef, Clef-Flash, and Clef-Omni: MLX MXFP4/MXFP8 and CUDA NVFP4 W4A16.
-
AutomatosX/AX-Clef-Flash-9B-CUDA-AXQ-NVFP4-W4A16
Text Generation • 9B • Updated • 25 -
AutomatosX/AX-Clef-27B-CUDA-AXQ-NVFP4-W4A16
Text Generation • 27B • Updated • 28 -
AutomatosX/AX-Clef-Flash-9B-MLX-AXQ-MXFP4
Text Generation • 9B • Updated • 30 -
AutomatosX/AX-Clef-Flash-9B-MLX-AXQ-MXFP8
Text Generation • 9B • Updated • 26
CUDA NVFP4 text, OCR, vision-language, and retrieval development checkpoints.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
AXQ Qwen3-Embedding and Nemotron-3-Embed retrieval packs: MLX MXFP4/MXFP8 and CUDA NVFP4.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
Qwen3.8: MLX MXFP4/MXFP8 and CUDA NVFP4 development packs with packaged MTP.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 71 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP8-MTP
Text Generation • 177B • Updated • 133
Qwen3-VL 4B/8B MLX MXFP4/MXFP8 plus CUDA NVFP4 W4A4. Vision BF16, no MTP.
-
AutomatosX/AX-Qwen3-VL-4B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 4B • Updated • 42 -
AutomatosX/AX-Qwen3-VL-8B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 9B • Updated • 42 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP4
Image-Text-to-Text • 4B • Updated • 67 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 4B • Updated • 72
DeepSeek-OCR-2 and Unlimited-OCR AXQ MX packs plus CUDA NVFP4 and six-bit lane.
-
AutomatosX/AX-DeepSeek-OCR-2-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 3B • Updated • 53 -
AutomatosX/AX-Unlimited-OCR-3B-MoE-CUDA-AXQ-NVFP4-W4A16
Image-Text-to-Text • 3B • Updated • 65 -
AutomatosX/AX-DeepSeek-OCR-2-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 85 -
AutomatosX/AX-DeepSeek-OCR-2-CUDA-AXQ-NVFP4-W4A16
Image-Text-to-Text • 3B • Updated • 73
MLX packs only. Native NVIDIA checkpoints are listed in the CUDA collection.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 71 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP8-MTP
Text Generation • 177B • Updated • 133
Source baselines and custom low-precision storage with explicit BF16/FP32 reconstruction. No native low-precision or quality certification.
AX coding packs: Tiel, Cyber-Tiel, Swift-1.5 Qwen3.8-27B, and Flash Next. MLX MXFP4 and MXFP8 with MTP. Not certified.
-
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 119 -
AutomatosX/AX-Swift-1.5-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 216 -
AutomatosX/AX-Swift-1.5-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 131
MLX MXFP8 checkpoints: OCR, Qwen3-VL, Qwen3.8 MTP, and embeddings.
-
AutomatosX/AX-DeepSeek-OCR-2-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 85 -
AutomatosX/AX-Unlimited-OCR-3B-MoE-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 151 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 4B • Updated • 72 -
AutomatosX/AX-Qwen3-VL-8B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 9B • Updated • 139
Hy-MT2 MLX MXFP4/MXFP8 and CUDA NVFP4 W4A16. Not certified.
AXQ packs with packaged MTP sidecars. -MTP ships assets; speed needs Tier 2.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k
Qwen3.8, Qwen3.5, VL, embeddings, and image reference storage.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 4B • Updated • 42
Nemotron-3-Embed 1B/8B retrieval and Super-120B/Lightning-30B MTP AXQ packs: MLX MXFP4/MXFP8 and CUDA NVFP4. Not certified.
-
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39 -
AutomatosX/AX-Nemotron-3-Embed-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 39 -
AutomatosX/AX-Nemotron-3-Super-120B-A12B-CUDA-AXQ-NVFP4-MTP
Text Generation • 124B • Updated • 483 -
AutomatosX/AX-Nemotron-3.5-Lightning-30B-A3B-CUDA-AXQ-NVFP4-MTP
Text Generation • 33B • Updated • 548
Tiel Coder 35B-A3B MXFP4/MXFP8 with packaged MTP, and the Cyber variant.
-
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 119 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 146
DeepSeek-OCR-2 AXQ MXFP4/MXFP8/NVFP4 and CUDA mixed six-bit.
Unlimited-OCR 3B MoE: native AXQ CUDA NVFP4, CUDA mixed six-bit, and MLX MXFP8/MXFP4.
Muse Glimmer NVFP4 W4A16 development conversion. Vision stays at source precision. Not quality-certified.
Source baselines and custom low-precision storage with explicit BF16/FP32 reconstruction. No native low-precision or quality certification.
Original AXDocLayout source and reference NVFP4/MXFP4/MXFP8 storage. Reference inference reconstructs FP32 PyTorch CPU weights.
AX coding packs: Tiel, Cyber-Tiel, Swift-1.5 Qwen3.8-27B, and Flash Next. MLX MXFP4 and MXFP8 with MTP. Not certified.
-
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 119 -
AutomatosX/AX-Swift-1.5-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 216 -
AutomatosX/AX-Swift-1.5-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 131
MLX AXQ MXFP4 development checkpoints; reference-only storage is listed separately.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k
MLX MXFP8 checkpoints: OCR, Qwen3-VL, Qwen3.8 MTP, and embeddings.
-
AutomatosX/AX-DeepSeek-OCR-2-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 85 -
AutomatosX/AX-Unlimited-OCR-3B-MoE-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 151 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 4B • Updated • 72 -
AutomatosX/AX-Qwen3-VL-8B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 9B • Updated • 139
AXQuant CUDA NVFP4 W4A16/W4A4 development containers. Runtime evidence is pack-specific.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
Hy-MT2 MLX MXFP4/MXFP8 and CUDA NVFP4 W4A16. Not certified.
Cloudflare Clef, Clef-Flash, and Clef-Omni: MLX MXFP4/MXFP8 and CUDA NVFP4 W4A16.
-
AutomatosX/AX-Clef-Flash-9B-CUDA-AXQ-NVFP4-W4A16
Text Generation • 9B • Updated • 25 -
AutomatosX/AX-Clef-27B-CUDA-AXQ-NVFP4-W4A16
Text Generation • 27B • Updated • 28 -
AutomatosX/AX-Clef-Flash-9B-MLX-AXQ-MXFP4
Text Generation • 9B • Updated • 30 -
AutomatosX/AX-Clef-Flash-9B-MLX-AXQ-MXFP8
Text Generation • 9B • Updated • 26
AXQ packs with packaged MTP sidecars. -MTP ships assets; speed needs Tier 2.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k
CUDA NVFP4 text, OCR, vision-language, and retrieval development checkpoints.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
Qwen3.8, Qwen3.5, VL, embeddings, and image reference storage.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 4B • Updated • 42
AXQ Qwen3-Embedding and Nemotron-3-Embed retrieval packs: MLX MXFP4/MXFP8 and CUDA NVFP4.
-
AutomatosX/AX-Qwen3-Embedding-0.6B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 0.6B • Updated • 45 -
AutomatosX/AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 4B • Updated • 39 -
AutomatosX/AX-Qwen3-Embedding-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 44 -
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39
Nemotron-3-Embed 1B/8B retrieval and Super-120B/Lightning-30B MTP AXQ packs: MLX MXFP4/MXFP8 and CUDA NVFP4. Not certified.
-
AutomatosX/AX-Nemotron-3-Embed-1B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 1B • Updated • 39 -
AutomatosX/AX-Nemotron-3-Embed-8B-CUDA-AXQ-NVFP4-W4A4
Sentence Similarity • 8B • Updated • 39 -
AutomatosX/AX-Nemotron-3-Super-120B-A12B-CUDA-AXQ-NVFP4-MTP
Text Generation • 124B • Updated • 483 -
AutomatosX/AX-Nemotron-3.5-Lightning-30B-A3B-CUDA-AXQ-NVFP4-MTP
Text Generation • 33B • Updated • 548
Qwen3.8: MLX MXFP4/MXFP8 and CUDA NVFP4 development packs with packaged MTP.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 71 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP8-MTP
Text Generation • 177B • Updated • 133
Tiel Coder 35B-A3B MXFP4/MXFP8 with packaged MTP, and the Cyber variant.
-
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.46k • 1 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP4-MTP
Text Generation • 35B • Updated • 1.24k -
AutomatosX/AX-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 119 -
AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Text Generation • 35B • Updated • 146
Qwen3-VL 4B/8B MLX MXFP4/MXFP8 plus CUDA NVFP4 W4A4. Vision BF16, no MTP.
-
AutomatosX/AX-Qwen3-VL-4B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 4B • Updated • 42 -
AutomatosX/AX-Qwen3-VL-8B-Instruct-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 9B • Updated • 42 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP4
Image-Text-to-Text • 4B • Updated • 67 -
AutomatosX/AX-Qwen3-VL-4B-Instruct-MLX-AXQ-MXFP8
Image-Text-to-Text • 4B • Updated • 72
DeepSeek-OCR-2 AXQ MXFP4/MXFP8/NVFP4 and CUDA mixed six-bit.
DeepSeek-OCR-2 and Unlimited-OCR AXQ MX packs plus CUDA NVFP4 and six-bit lane.
-
AutomatosX/AX-DeepSeek-OCR-2-CUDA-AXQ-NVFP4-W4A4
Image-Text-to-Text • 3B • Updated • 53 -
AutomatosX/AX-Unlimited-OCR-3B-MoE-CUDA-AXQ-NVFP4-W4A16
Image-Text-to-Text • 3B • Updated • 65 -
AutomatosX/AX-DeepSeek-OCR-2-MLX-AXQ-MXFP8
Image-Text-to-Text • 3B • Updated • 85 -
AutomatosX/AX-DeepSeek-OCR-2-CUDA-AXQ-NVFP4-W4A16
Image-Text-to-Text • 3B • Updated • 73
Unlimited-OCR 3B MoE: native AXQ CUDA NVFP4, CUDA mixed six-bit, and MLX MXFP8/MXFP4.
MLX packs only. Native NVIDIA checkpoints are listed in the CUDA collection.
-
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP4-MTP
Text Generation • 27B • Updated • 903 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP4-MTP
Text Generation • 177B • Updated • 677 -
AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-MXFP8-MTP
Text Generation • 27B • Updated • 71 -
AutomatosX/AX-Qwen3.8-Flash-Next-MLX-AXQ-MXFP8-MTP
Text Generation • 177B • Updated • 133