ResearchBenchmarksMathematical Problem Solving on MathArena FA TotalMathematical Problem Solving on MathArena FA TotalFollow60Correctness RatioCIKA5758.56061.5May 8, 2026Evaluation resultsMethodMethodLinksCorrectness RatioAccuracyCIKAsampling=pass@1, shots...sampling=pass@1, shots=1-shot2026.056046.2