ResearchBenchmarksMath Reasoning on HMMT 25 (Feb)Follow22.29Pass@1XRPO17.0918.4419.7921.14Oct 8, 2025Evaluation ResultsMethodMethodLinksPass@1Cons@32XRPOBackbone=Qwen3-1.7BBackbone=Qwen3-1.7B2025.1022.2926.67GSPOBackbone=Qwen3-1.7BBackbone=Qwen3-1.7B2025.1021.2520DAPOBackbone=Qwen3-1.7BBackbone=Qwen3-1.7B2025.1020.4223.33TPO-SBackbone=Qwen3-1.7BBackbone=Qwen3-1.7B2025.1017.2920