ferrumox/rabbit
Run Qwen 3.8 MAX (2.4T MoE) on a 50GB-RAM consumer machine in pure Rust, experts streamed from disk. Tiny engine, immense model.
4 best points · 0 comments · 1 Hacker News post · 192 GitHub stars · View on GitHub
Hacker News discussions
- Ran Moonshot's 2.8T-parameter Kimi K3 on a GPU-less mini-PC (4 points, 0 comments, by manuelslemos, 2026-07-29)