Revision History
GAIA benchmark · 8 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: Corrected GAIA leaderboard figures: removed a non-existent 92.36% openJiuwen entry and fixed its real score (91.69%), relabeled the 92%+ scores as test-set (not validation) results, fixed Manus and Genspark figures that were Level 1 scores presented as overall, corrected Deep Research's average to 67.36%, AutoGPT's overall to about 5%, the 2024 Hugging Face 44.2% result (Transformers Agents, July 2024), the test-set Level 2/3 question counts (159/49), the annotation-validation process, and removed unsupported claims about a ChatGPT Agent GAIA result, retired questions, an extended scorer, and other errors (plus 28 minor fixes)
Version 8 · Nov 5, 2026, 03:09 AM
- Improve internal links (2026-10-13)
Version 7 · Oct 13, 2026, 10:15 AM
- Added 6 contextual internal links
Version 6 · Jul 23, 2026, 03:27 PM