Trending storyDeveloping
Apple demos 816B model running locally on a Mac cluster
1 report1 sourceupdated 2 hours ago
What happened
AI digest
An October 15 report says Apple demonstrated running an 816B model locally on a cluster of Mac Studios. The demo used four M5 Ultra Mac Studios with MLX and Thunderbolt RDMA, running the Muse Spark Open 816B model. The reported result: 36.73 tokens per second at an 8,192-token context. So far the report covers only the demo's hardware setup, the technology used, the model and the performance figure.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
Oct 16
- AI HOT · Tips & opinionsApple 演示用 MLX 与 RDMA 在 Mac Studio 集群本地运行 816B 大模型
Apple 演示用 MLX 和 Thunderbolt RDMA 在四台 M5 Ultra Mac Studio 上运行 Muse Spark Open 816B，在 8,192 tokens 上下文下达到 36.73 tokens per second。
Heat over time
Not enough continuous observations to draw a trend yet.