WorldArena Track 2 - Policy Evaluator: leaderboard

WorldArena Track 2 policy-evaluator score measuring how well world-model outputs support policy evaluation for embodied agents.

Metric: Policy Evaluator Score (%). Source: huggingface.co. 19 models tracked.

Top models

#ModelScoreOverall rank
1WorldScape v0.299.53
2chill99.22
3Ctrl-World98.6
4DSCFuncWorld98.54
5Curr_World98.47
6BLM97.78
7upset97.75
8WoW95.88
9py95.36
10N-World94.19

Interactive version: theaggregate.ai/benchmark?slug=worldarena-track-2-policy-evaluator · How It Works · Data refreshed daily, snapshot 2026-10-20.