Reasoning Arena: Trace Tournaments When Verifiable Rewards Fall Short
About
Reinforcement learning with verifiable rewards (RLVR) has become a leading paradigm for improving the reasoning ability of large language models through outcome-based supervision. However, verifiable rewards frequently become uninformative at the group level: when all sampled traces of a given prompt receive identical rewards, group-relative advantage estimation provides no gradient signal, even though the traces may differ substantially in reasoning quality. We propose Reasoning Arena, an adaptive training framework that routes such non-diverse reward groups to a judge system instead of discarding them. Beyond examining the final answer, Reasoning Arena constructs trace tournaments, where reasoning traces are compared head-to-head to expose finer-grained preferences within the group, converting reasoning quality into rich relative reward signals. To make reward estimation efficient, rather than exhaustively comparing every pair, each new trace is evaluated against a small, dynamically updated pool of previously generated traces as anchors to efficiently establish a relative ranking. We then fit a Bradley-Terry model on the incomplete comparison graph, enabling scalable RL integration without quadratic pairwise comparisons. Empirical results demonstrate that Reasoning Arena consistently outperforms the RLVR baseline by 7.6% on average in competition mathematics and coding benchmarks. By converting otherwise wasted zero-advantage samples into useful gradient updates, our method accelerates training by 27% to 41%, saving nearly 50% of generation compute, and substantially improves overall reasoning performance.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Code Generation | LiveCodeBench v6 | Pass Ratio52.2 | 6 | |
| Competition Mathematics | AIME 2024 | Pass Ratio66.7 | 6 | |
| Competition Mathematics | AIME 2025 | Pass Ratio54.8 | 6 | |
| Competition Mathematics | AIME 2026 | Pass Ratio59 | 6 | |
| Competition Mathematics | Beyond AIME | Pass Ratio36.4 | 6 | |
| Scientific Reasoning | GPQA Diamond | Pass Ratio60.5 | 6 |