How HumanEval scored a B.

A way.space signal summary built from public evidence. It is not a user rating, and not a safety certification. How scores are calculated.

01 / Breakdown

Derived from GitHub, not voted.

70/100

GitHub credibility score

code
45/100
- ✓ Contributors & bus-factor: 5 recent contributors
- ✓ Commit recency: last commit 6d ago
- ✓ Commit cadence: 2 commits in 90 days
- ✓ Issue close ratio: 22% of issues closed
- ↓ Release cadence: 0 releases in 12 months
- ✓ Stars (log-damped): 3,414 stars (log-damped)
- – Star authenticity: star-velocity check unavailable — GitHub no longer publishes stargazer timestamps for this repo
- – Dependent repos: not measurable — no package listing on a registry we can query
identity
80/100
- ✓ Organisation: Built by OpenAI · notable organization (verified: Org profile shows a Verified badge stating the organization openai controls the domain openai.com; profile website https://openai.com/. No description or location field shown.)
audits
Not measured
- – Third-party scans: not measurable — no registry listings for this tool
freshness
100/100
- ✓ Last shipped: last commit 6d ago — standard-churn tools are expected to ship about every 120d
community
95/100
- – Recommends: no recommends from gate-passing builders yet
- ✓ Hacker News mentions: 17 Hacker News stories pointing at this tool

Components that could not be measured are left out of the average entirely, and their weight is redistributed, never scored as zero. Identity and Community are left out the same way when they would not have raised the score. How scores are calculated.