SpatialWorld - Daily Routines: leaderboard

Metric: Task success rate (%) on the 350 daily-routine tasks (AI2-THOR, ProcTHOR, VirtualHome, EmbodiedCity); task success rate (%) judged by terminal-state verifiers, vision-only egocentric observation and a unified text action interface, temperature 1.0, 30-turn history, step budget twice the human golden action count plus 10; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 15 models tracked.

Top models

#ModelScore
1GPT-514.9
2Qwen 3.5 397B A17B13.1
3Gemini 3.1 Pro (Preview)11.4
4Kimi K2.511.1
5GPT-5.48
6Gemini 3 Flash8
7Gemini 2.5 Pro7.4
8Qwen 3 VL 235B A22B Instruct6.9
9Qwen 3 VL 235B A22B (Thinking)5.7
10Seed 2.0 Lite5.7
11Qwen 2.5 VL 72B Instruct3.7
12GLM-4.5V3.7
13GLM-4.6V2.9

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