WeaveBench - Spatial and 3D: leaderboard

Metric: PassRate (%) on the Spatial/3D domain tasks of 114 long-horizon real-world computer-use tasks that need GUI, CLI, code and API tools together, each model run as the backbone of a fixed OpenClaw agent runtime with the full hybrid tool pool at its best thinking mode; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 10 models tracked.

Top models

#ModelScore
1GPT-5.5 (High)33.3
2Claude Opus 4.7 (High)16.7
3Gemini 3.1 Pro (Preview)8.3
4GPT-5.4 (High)8.3
5GPT-5.3 Codex (High)8.3
6Qwen 3.5 397B A17B0
7Qwen 3 VL 8B (Thinking)0
8GPT-5.1 Codex (High)0

Interactive version: theaggregate.ai/benchmark?slug=weavebench-spatial-and-3d · How It Works · Data refreshed daily, snapshot 2026-09-29.