Plan-RewardBench - Error Recovery: leaderboard
Plan-RewardBench (2026) tests reward models and LLM judges on preferring the better of two tool-using agent trajectories, with preference labels from a judge panel audited by humans. This board is pairwise accuracy in percent on the pairs that test recovery from tool errors; higher is better.
Metric: Pairwise accuracy (%) on the 361 robust error-recovery pairs of the Plan-RewardBench chosen versus rejected agent trajectories (shared tool registry, multi-turn context, executed tool feedback), preference labels from a Gemini 3 Pro and GPT-5.1 judge panel with human audit; discriminative reward models score each trajectory, generative reward models and LLM judges choose between the pair; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 21 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Skywork-Reward-V2-Qwen3-8B | 75.35 | |
| 2 | Qwen 3 30B A3B 2507 Instruct | 74.93 | #1100 |
| 3 | Qwen Plus | 73.75 | #820 |
| 4 | INF-ORM-Llama3.1-70B | 69.78 | |
| 5 | GPT-5 | 69.39 | #178 |
| 6 | Qwen 3 235B A22B 2507 Instruct | 68.84 | #616 |
| 7 | Qwen 3 4B 2507 Instruct | 68.7 | #2221 |
| 8 | Qwen Max | 68.63 | #1092 |
| 9 | DeepSeek R1 | 68.01 | #482 |
| 10 | Gemini 3 Flash (Preview) | 67.31 | #152 |
| 11 | RRM-32B | 67.15 | #1115 |
| 12 | InternLM2-7B-Reward | 67.05 | |
| 13 | DeepSeek V3.2 Exp | 66.76 | #439 |
| 14 | Qwen 3 235B A22B 2507 (Thinking) | 66.48 | #466 |
| 15 | Skywork-Reward-V2-Llama-3.1-8B | 65.1 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=plan-rewardbench-error-recovery · How It Works · Data refreshed daily, snapshot 2026-11-21.