GDPval leaderboard
GDPval
2 models tested · Updated 2026-04-23 · Verified sources only
GPT-5.5 leads at 84.9%
1
GPT-5.5
OpenAI · Blog/OpenAI · 2026-04-23
GDPval measures knowledge-work quality across 44 occupations. New SOTA.
84.9%
2
GPT-5.4
OpenAI · Blog/OpenAI · 2026-03-05
Real-world knowledge work benchmark: 1,320 tasks across 44 occupations in 9 GDP sectors. Frontier models approach expert quality at 100x speed/cost.
83.0%
benchmark.space canonical public ledger of AI capability