Artificial Analysis Posts GDP.pdf Model Leaderboard
Artificial Analysis shares results from the GDP.pdf document reasoning benchmark created by HelloSurgeAI.
TLDR
Artificial Analysis posted results from the GDP.pdf evaluation on X. The benchmark created by HelloSurgeAI tests single-turn professional document reasoning across 100 PDFs in ten domains that together contain 4,592 pages. The company states that OpenAI leads with GPT-6 Astra at 33.2% and GPT-5.6 Sol at 28.2%, followed by Claude Fable 5.1 at 26.2%. The posts include two-panel charts and a grid of bar charts titled Intelligence Evaluations.
Combined views
37.8K
2 Sources, first seen 26d ago