Repository Public
1{2 "models": [3 {4 "id": "maclustr-qwen3",5 "display_name": "Qwen3-30B-A3B Instruct (MLX 4-bit)",6 "engine_model": "~/maclustr-llm-data/models/qwen3-30b-a3b-instruct-2507-4bit",7 "repo": "mlx-community/Qwen3-30B-A3B-Instruct-2507-4bit",8 "revision": "e9675aa3ca5f900ccef55267914466d55ab325fa",9 "quantization": "MLX 4-bit",10 "license": "apache-2.0",11 "context_window": 32768,12 "max_output_tokens": 8192,13 "default_max_tokens": 1024,14 "supports_tools": true,15 "supports_thinking": false,16 "description": "Mixture-of-Experts 30,5 G de paramètres, 3,3 G actifs par token. Servi localement par Apple MLX sur Mac mini M4 Pro.",17 "created": 179504640018 }19 ],20 "engines": [21 { "id": "m4mm", "url": "http://127.0.0.1:8811", "models": ["maclustr-qwen3"], "max_active": 8, "token_budget": 48000 }22 ]23}24