Newborn Newborn C++ C+ CU
61
q27 signalnine/q27

Quasar is an inference engine for Qwen3.6-27B-MTP / Qwen3.8-27B on a single RTX 5090, in the spirit of antirez/ds4

HEALTH 100 / Healthy
// solo builder// hidden gem// release machine

// readme

Quasar

A narrow inference engine for Qwen3.6-27B-MTP and Qwen3.8-27B-MTP (hybrid GDN+attention, trained-in MTP heads) and their fine-tunes on a single RTX 5090 (3090 and 4090/Ada also supported; Apple-silicon Metal backend for the q4s tier). One model family, one GPU, as fast as possible. In the spirit of antirez/ds4

Why this is interesting

  • Fastest of the four engines tested, on this harness — across two independent runs. Mean wall per SWE-bench instance, same 12 tasks, one harness, unchanged competitor binaries:

    leg2026-08-172026-08-19
    q27 q5f46.8 s46.3 s
    q27 q4s48.5 s49.6 s
    llama.cpp71.4 s59.9 s
    vLLM84.5 s78.8 s
    ninfer NVFP496.8 s113.8 s
    ninfer int8327.4 s266.2 s

    Two runs is what it takes to see the honest shape: q27 is the only leg that reproduced (46.8 -> 46.3 s), while every competitor moved 7-19% on identical binaries. So the ordering is stable but the margins are not — q27’s edge over llama.cpp reads 1.53x one day and 1.29x the next. Quoting either as the number overstates the precision.…

The Undervalued Score +

How much a project earns versus how much attention it actually gets. Above 50 means the work is outrunning its audience. Recomputed nightly from commit velocity, contributor effort, issue resolution, fork utility, release cadence, and project maturity — divided by a logarithmic reach factor.

score  = signal / reach

signal = 0.25·commit_velocity   // commits in last 90 days (cap 30)
       + 0.20·contributor_work  // unique authors × velocity (cap 100)
       + 0.20·issue_resolution  // closed ÷ total issues
       + 0.20·fork_ratio        // forks ÷ stars (proxy for real usage)
       + 0.10·release_cadence   // releases in 90 days (cap 3)
       + age_bonus              // +0 to +0.30 after 6 months
       + homepage_bonus         // +0.05 if homepage is set

reach  = log₁₀(stars + watchers + 10)
The Health Score +

Is the project alive and maintained right now? A 0–100 pulse recomputed nightly from commit recency, rhythm, how fast issues close, and how quickly PRs get merged.

health = 0.35·recency       // days since last commit (90d decay)
       + 0.25·cadence       // commit rhythm consistency
       + 0.20·issue_health  // closed ÷ total issues
       + 0.20·pr_health     // merged ÷ total PRs
Health bands +

The colour and label on every card come straight from the health score.

Healthy   80 – 100   active, responsive, regular releases
Stable    60 – 79    maintained, steady, no alarms
Quiet     40 – 59    slowing down — watch this one
At Risk    0 – 39    going dark · candidate for rescue
// Tags — what each label means +

Tags are independent behavioral signals computed nightly. A project can hold multiple at once. They drive the home page sections.

solo_builder      one person holds > 80% of commits (last 180d)
needs_contributors has open "help wanted" or "good first issue" labels
hidden_gem        < 100 stars · active in last 3 months · documented
legacy_hero       repo > 5 years old · committed this year
fork_magnet       forks/stars > 0.5 · used as template or dependency
release_machine   5+ releases in the last 90 days
under_pressure    > 10 open issues · ≤ 2 contributors · health ≥ 60
community_watch   watchers > stars · devs tracking before the public
community_hub     GitHub Discussions enabled · > 20 discussions
funded            maintainer has active funding channel
Why rank against stars at all? +

Stars are an outcome, not effort. A project with 8 stars and daily commits is doing more interesting work than one coasting on 8k. We measure the building, then divide by the attention already received — so the genuinely undervalued rise to the top.

// stars   = lagging indicator
// commits = leading indicator
// we rank by the leading one