ResearchBenchmarksMath reasoning on MATH 500Math reasoning on MATH 500Follow3.95Tau (Mean Accepted Tokens)LongSpec0.8821.67852.4753.2715Feb 24, 2025Evaluation resultsMethodMethodLinksTau (Mean Accepted Tokens)Throughput (Tokens/s)Improvement FactorLongSpecModel=QwQ-32B, Maximum...Model=QwQ-32B, Maximum output length=32k tokens2025.023.9548.362.47VanillaModel=QwQ-32B, Maximum...Model=QwQ-32B, Maximum output length=32k tokens2025.02119.59—