WM Bench

World Model Bench leaderboard for cognitive intelligence in world models and embodied AI systems, scoring perception, cognition, embodiment, and ten scenario categories.

13rows
wm_scoreprimary metric
2026-05-06sampled

Metadata

ID
wm_bench
Category
Embodied AI
Release
Unknown

Metrics

WM Score, P1 Perception, P2 Cognition, P3 Embodiment, FPS, Cognitive Latency (lower is better), C01 Environmental Awareness, C02 Entity Recognition, C03 Prediction-Based Reasoning, C04 Threat-Type Differentiation, C05 Autonomous Emotion Escalation, C06 Contextual Memory Utilization, C07 Post-Threat Adaptive Recovery, C08 Motion-Emotion Expression, C09 Real-Time Cognitive Performance, C10 Body-Swap Extensibility

Latest Results

Rows are parsed from the public WM Bench leaderboard's static LB_DATA block. The source's estimated versus verified status is preserved in metadata.source_estimated.

Rank Subject WM Score Model Match Provenance Sampled
1 PROMETHEUS v1.0 726 — Imported 2026-05-06
2 Meta V-JEPA 2-AC 554 — Imported 2026-05-06
3 Wayve GAIA-3 550 — Imported 2026-05-06
4 NC AI WFM v1.0 522 — Imported 2026-05-06
5 NVIDIA Cosmos v1.0 498 — Imported 2026-05-06
6 NAVER LABS SWM 470 — Imported 2026-05-06
7 DeepMind Genie 2 449 — Imported 2026-05-06
8 DreamerV3 XL 441 — Imported 2026-05-06
9 OpenAI Sora 2 381 — Imported 2026-05-06
10 World Labs Marble 362 — Imported 2026-05-06
11 UniSim 338 — Imported 2026-05-06
12 DIAMOND v1.0 312 — Imported 2026-05-06
13 Oasis AI 285 — Imported 2026-05-06