Apple

Latest AI news, models and releases from Apple. ['A19 Pro', 'Applebot-Extended', 'Apple Health', 'Apple Intelligence', 'Foundation Models framework', 'M3 Ultra', 'M5 Max', 'M5 Pro', 'M5 Ultra', 'M6', 'M7 Max', 'M7 Pro', 'Mac Studio', 'MLX', 'Qwen3-VL-8B', 'Siri']

Mac Mini M4 Pro and OpenClaw: Testing Local LLM Inference

The article tests local LLM inference on a Mac mini M4 Pro with 48 GB RAM using OpenClaw and LM Studio. Three models are compared: google/gemma-4-26b-a4b-qat, qwen/qwen3.6-35b-a3b, and google/gemma-4-31b. Results show that MoE models like gemma-4-26b and qwen3.6-35b offer better speed and energy efficiency compared to the dense model gemma-4-31b.

AppleApple OpenAIOpenAI Google/DeepMindGoogle/DeepMind
Habr — хаб ИИ24.07 · 03:01
Fresh news