Skip to content

chore: cleanup duplicate GPT 5.2 Codex leaderboard entries#55

Open
Chesars wants to merge 1 commit intoSWE-bench:masterfrom
Chesars:fix/remove-duplicate-gpt52-codex
Open

chore: cleanup duplicate GPT 5.2 Codex leaderboard entries#55
Chesars wants to merge 1 commit intoSWE-bench:masterfrom
Chesars:fix/remove-duplicate-gpt52-codex

Conversation

@Chesars
Copy link
Contributor

@Chesars Chesars commented Mar 3, 2026

Summary

  • Removes duplicate "GPT 5.2 Codex" entries from bash-only and multilingual leaderboards
  • PR Add GPT 5.2 Codex (high reasoning) and fix GPT 5.2 naming #50 manually added these entries to leaderboards.json, but the same model already existed as "GPT-5-2 Codex" (auto-generated from experiments repo)
  • The duplicates had logs: null while the canonical entries have proper S3 log paths

Details

Leaderboard Kept (from pipeline) Removed (manual duplicate)
bash-only GPT-5-2 Codex (72.8%, has logs) GPT 5.2 Codex (72.8%, logs=null)
Multilingual GPT-5-2 Codex (66.3%, has logs) GPT 5.2 Codex (66.3%, logs=null)
Screenshot 2026-03-03 at 08 34 36

…l leaderboards

PR SWE-bench#50 manually added "GPT 5.2 Codex" entries to leaderboards.json, but
the same model already existed as "GPT-5-2 Codex" (auto-generated from
the experiments repo). This removes the manual duplicates (logs=null)
and keeps the canonical entries from the pipeline.
@Chesars Chesars changed the title Remove duplicate GPT 5.2 Codex leaderboard entries chore: Cleanup duplicate GPT 5.2 Codex leaderboard entries Mar 3, 2026
@Chesars Chesars changed the title chore: Cleanup duplicate GPT 5.2 Codex leaderboard entries chore: cleanup duplicate GPT 5.2 Codex leaderboard entries Mar 3, 2026
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