ResearchBenchmarksData Science on Scaling Law Discovery parallelFollow1R2 ScoreSLDAgent0.950.97511.025Apr 21, 2026Evaluation ResultsMethodMethodLinksR2 ScoreSLDAgentModel=GPT-5Model=GPT-52026.041SIMPLETESModel=gpt-oss-120bModel=gpt-oss-120b2026.041