{"status": "ok", "model": "Qwen3.8-27B-MLX-4bit", "model_ids": ["Qwen3.8-27B-MLX-4bit"], "max_batch_size": 1, "warming": false, "memory": {"active": 15558864464, "cache": 0, "peak": 15559077456}, "live": {"connections": 0, "waiting": 0, "decode_tokens_per_second": 0.0, "prefill_tokens_per_second": 0.0}, "version": "1.2.3", "backend": "metal", "device": "Apple M4 Max"}