# Local AI Release Board
> Tracks new open-weight model releases (Hugging Face) and local-AI runtime/harness releases (GitHub), with an estimated "runs on" mapping to consumer GPUs, DGX Spark, Macs and CPU. Window: last 30 days. Generated 2026-10-15T18:13:06Z.

- [Board](https://localai-board.pages.dev/): filterable list
- [JSON feed](https://localai-board.pages.dev/api/releases.json): models[], harnesses[], ollama_newest[], x_posts[], devices[], estimate_note
## Notes
- Facts come from the HF API/model cards, GitHub releases, ollama.com/library and public X posts. Unknown fields are left blank.
- "Runs on" values are estimates: ESTIMATE ONLY. Need = weights at ~4.8/8.5/16 bits per weight + FP16 KV cache for 8192 tokens (from config.json when available, otherwise a rough 8% of Q8 weights) + 1.0GB overhead. Usable memory per device is assumed below physical (Mac ≈75% of RAM for GPU). Arc classes only count Q8/FP16, matching what performs well on Arc. MoE models still need all weights resident; active params only affect speed. Not verified by running.