Trending storyDeveloping
Mac LLM calculator estimates whether a Mac can run a model and how fast
1 report1 sourceupdated 2 hours ago
What happened
AI digest
A Mac LLM calculator posted to the Hacker News front page on October 17 estimates whether a given Mac can run a model and how fast it will generate. It works from model weights, unified memory and memory bandwidth, matching a model's needs to the Mac's hardware. One example: an M4 Pro with 64 GB running Q4_K_M Llama 3.3 70B at about 5.7 tokens per second. That figure is the calculator's estimate, not a measured benchmark.
Written by AI from the coverage · updated 37 minutes ago
Coverage
Follow the reports to see the story from different sides.
Oct 17
- Hacker News front pageShow HN: Which Mac runs which LLM, and how fast
Mac LLM 计算器根据模型权重、统一内存和内存带宽，估算不同 Mac 能否运行指定模型及其生成速度。计算器预计，64 GB 的 M4 Pro 可运行 Q4_K_M 量化的 Llama 3.3 70B，生成速度约为 5.7 tokens/sec。
Heat over time
Not enough continuous observations to draw a trend yet.