ResearchBenchmarksPaper Understanding on ELAIPBenchFollow43.7ScoreAgentSPEX33.40436.07738.7541.423Apr 14, 2026Evaluation ResultsMethodMethodLinksScoreAgentSPEXModel=GPT-5*, Domain=P...Model=GPT-5*, Domain=Paper Understanding2026.0443.7CoTModel=GPT-5*, Domain=P...Model=GPT-5*, Domain=Paper Understanding2026.0437.22ReActModel=GPT-5*, Domain=P...Model=GPT-5*, Domain=Paper Understanding2026.0433.8