Skip to content

Pull requests: openai/frontier-evals

New pull request New
41 Open 87 Closed
41 Open 87 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Make nanoeval global concurrency check atomic
#171 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix SWE-Lancer default runs directory
#170 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Honor zero-hour PaperBench checkpoint selection
#169 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Apply default nanoeval recorder sample context
#168 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Propagate external cancellation from EVMbench agent runs
#167 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Limit PaperBench computer retries to startup
#166 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix PaperBench zero prior-node limit
#165 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix PaperBench code-only fallback for non-code rubrics
#164 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Preserve nanoeval max concurrency when resuming run sets
#163 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Group binary agent outcomes for standard metrics
#162 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Handle empty EVMbench summaries
#161 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix PaperBench rubric pruning at subtree boundaries
#160 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Reject zero nanoeval multiprocessing workers
#159 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Normalize EVMbench scores at the nanoeval boundary
#158 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Use configured EVMbench incorrect finding text
#157 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Tighten compatibility Docker registry validation
#156 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix EVMbench medium findings subdirectory option
#155 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Fix multi-answer MCQ consensus group collisions
#154 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
Treat invalid SWE-Lancer pytest runs as rollout errors
#153 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
1
Fix MCQ API answer-format prompt
#152 opened Aug 17, 2026 by sylvesterkaczmarek Loading…
1
Preserve PaperBench run group identity when aggregating seeds
#150 opened Aug 16, 2026 by sylvesterkaczmarek Loading…
2
6
Fix SWE-Lancer token usage totals
#148 opened Aug 16, 2026 by sylvesterkaczmarek Loading…
1
Skip null PaperBench token usage during aggregation
#147 opened Aug 16, 2026 by sylvesterkaczmarek Loading…
1
Propagate PaperBench judge validity through aggregate nodes
#145 opened Aug 16, 2026 by sylvesterkaczmarek Loading…
1
Load legacy PaperBench reproduction metadata with optional defaults
#143 opened Aug 16, 2026 by sylvesterkaczmarek Loading…
1
ProTip! Filter pull requests by the default branch with base:main.