ResearchBenchmarksMath Reasoning on AIME 2025 edition (test)Math Reasoning on AIME 2025 edition (test)Follow4,377Optimization TokensETGPO3,978.86,666.659,354.512,042.35Feb 1, 2026Evaluation resultsMethodMethodLinksOptimization TokensETGPOBackbone model=GPT-4.1...Backbone model=GPT-4.1-mini, Optimizer=GPT-4.12026.024,377GEPABackbone model=GPT-4.1...Backbone model=GPT-4.1-mini, Optimizer=GPT-4.12026.0212,495MIPROv2Backbone model=GPT-4.1...Backbone model=GPT-4.1-mini, Optimizer=GPT-4.12026.0214,332