EMem-Bench - Passive Observation: leaderboard

Metric: Success rate (%; the 1036 Passive Observation episodes, finding an object seen earlier in an object-dense scene; full-context condition: the complete interaction history as sanitized text plus the current 500 x 500 RGB view, one primitive action per call with deterministic decoding and at most eight target-task steps in AI2-THOR and ProcTHOR homes; an episode succeeds only if the simulator reaches the target terminal state). Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.

Top models

#ModelScore
1Gemini 3 Flash67
2Gemini 2.5 Pro63.5
3Qwen 3.6 27B61
4GPT-5.460.4
5GPT-5.4 Mini55.1
6InternVL3-38B47.9
7Qwen 3 VL 32B Instruct47.1
8Qwen 3 VL 4B Instruct46.4
9Qwen 3 VL 8B Instruct41.4
10InternVL3-8B20.5

Interactive version: theaggregate.ai/benchmark?slug=emem-bench-passive-observation · How It Works · Data refreshed daily, snapshot 2026-09-26.