ResearchBenchmarksMathematical Theorem Proving on AIME 2024Follow6.9Proof-level AccuracyBASE3.4684.3595.256.141Jun 14, 2026Evaluation ResultsMethodMethodLinksProof-level AccuracyBASESolver=Qwen3-8B, Evalu...Solver=Qwen3-8B, Evaluation Protocol=BASE, Prover=DeepSeek-Prover-V2, K=82026.066.9BASESolver=Gemini 3 Pro, E...Solver=Gemini 3 Pro, Evaluation Protocol=BASE, Prover=DeepSeek-Prover-V2, K=82026.066.7BASESolver=R1-Qwen3-8B, Ev...Solver=R1-Qwen3-8B, Evaluation Protocol=BASE, Prover=DeepSeek-Prover-V2, K=82026.063.6