PlanBench-V - Scheme Evaluation: leaderboard

Metric: Mean score (out of 2) on judging the strengths and weaknesses of a planning proposal, implementation level; answers to expert-curated questions on Chinese territorial spatial planning maps are scored 0 to 2 by a GPT-4o-mini judge (temperature 0) against structured reference answers with annotated critical points; objective items use exact match or semantic similarity; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.

Top models

#ModelScore
1Qwen 3.6 Plus1.62
2GPT-5.41.59
3Kimi K2.61.44
4Gemini 2.5 Pro1.44
5Claude Opus 4.71.43
6GPT-4o1.22
7Qwen 3.6 Flash1.05
8Qwen 2 VL 7B0.86
9Qwen 2.5 VL 7B Instruct0.8
10InternVL3-14B0.71
11GPT-4o Mini0.64
12InternVL3-8B0.63
13Qwen 2 VL 2B0.54

Interactive version: theaggregate.ai/benchmark?slug=planbench-v-scheme-evaluation · How It Works · Data refreshed daily, snapshot 2026-09-29.