Plan-RewardBench: 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, the macro-average over the seven scenarios; higher is better.
Metric: Pairwise accuracy (%), macro-average of the seven scenario columns 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 | Qwen Plus | 69.96 | #828 |
| 2 | DeepSeek V3.2 Exp | 69.57 | #440 |
| 3 | Qwen 3 235B A22B 2507 Instruct | 69.54 | #625 |
| 4 | INF-ORM-Llama3.1-70B | 69.21 | |
| 5 | Gemini 3 Flash (Preview) | 69.07 | #156 |
| 6 | DeepSeek R1 | 68.95 | #484 |
| 7 | Qwen Max | 68.92 | #1094 |
| 8 | Qwen 3 235B A22B 2507 (Thinking) | 68.6 | #471 |
| 9 | GPT-5 | 68.54 | #185 |
| 10 | Qwen 3 4B 2507 Instruct | 67.4 | #2240 |
| 11 | InternLM2-7B-Reward | 66.96 | |
| 12 | Qwen 3 30B A3B 2507 Instruct | 66.75 | #1106 |
| 13 | Skywork-Reward-V2-Qwen3-8B | 66.46 | |
| 14 | RRM-32B | 66.45 | #1127 |
| 15 | Skywork-Reward-V2-Llama-3.1-8B | 65.88 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=plan-rewardbench · How It Works · Data refreshed daily, snapshot 2026-11-22.