Skip to content

0.2.2: README leads with regraded SWE-bench Verified result; CHANGELOG dates - #6

Merged
surlymochan merged 2 commits into
mainfrom
ship/0.2.2
Oct 17, 2026
Merged

surlymochan merged 2 commits into
mainfrom
ship/0.2.2

Conversation

@surlymochan

Copy link
Copy Markdown
Owner
  • README: SWE-bench Verified block corrected after regrade (per-run TMPDIR; pytest temp-dir race): openswarm 5/6, OpenFlow 4/6, baseline 3/6, superpowers 2/6 on DeepSeek-V4.1-Flash, n=1. Caveats on n=1 and ~1.8x time / +61-76% $ per task; link to raw data and Optimization backlog from SWE-bench Verified results (overhead, measurement, next round) #5.
  • bench/swebench: regrade.py, apply_regrade.py, regraded records, Regrade section; >5 MB diff stored as .gz + sha256.
  • CHANGELOG: 0.2.0 / 0.2.1 dated 2026-10-17; 0.2.2 entry. RELEASE_NOTES 0.2.2. package.json 0.2.2.
    No CLI or gate changes.

@surlymochan
surlymochan merged commit 86cc9f9 into main Oct 17, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant