The fastest way to run Qwen 3.8 Flash Next and Qwen 3.8 27B on a Mac: 125 tok/s in OpenCode on an M5 Max. Native MTP speculative decoding on Apple Silicon, exact at any temperature. OpenAI and Anthropic compatible local server.
youssofal/MTPLX is drawing steady momentum at +4.0 stars/day (55th percentile in the tracked cohort), for a momentum score of 34.2/100. A recent release is helping fuel the run.
Low breakout odds over the next 14 days, led by push recency (87th pctl) and star velocity (55th pctl). Confidence is high given the available history.
Reconstructed from 60 snapshots — the time-series GitHub’s API doesn’t expose.
Release count rose to 56.
Release count rose to 54.
Live score, refreshed every 6 hours. Links back to this page.
[](https://repoclimb.com/repo/youssofal/MTPLX)Transparent heuristic · logged for model training
Percentile rank within the tracked cohort. The score self-calibrates — it answers “accelerating vs. everything else,” not raw size.