EgoIntrospect - Memory Recall Prediction: leaderboard
Which of four details of a three-minute clip the wearer later recalled as vivid, asked on egocentric smart-glasses clips annotated by the wearers themselves (EgoIntrospect, 2026); native audio and video models hear the clip audio and the others read a transcript. Accuracy in percent on 200 four-choice questions; higher is better.
Metric: Accuracy (%) on 200 four-choice questions asking which detail of a 3-minute clip the wearer later recalled as vivid (distractors are other elements of the same clip written by Gemini and curated by the authors), fixation-overlaid egocentric clips from 15 held-out participants with the participant profile; native audio-video models hear the clip audio, the other models read its transcript; temperature 0 (0.6 for InternVL3.5 thinking mode); higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 19 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Qwen3 Omni 30B A3B Instruct | 55.5 | #930 | |
| 2 | Qwen 3.5 Omni Plus | 52 | #254 | |
| 3 | GPT-4o | 51.5 | #751 | |
| 4 | Gemini 3.1 Pro (Preview) | 50 | #99 | |
| 5 | MiniCPM-V-4.5 | 50 | #1441 | |
| 6 | Qwen 3.5 397B A17B | 49 | #264 | |
| 7 | LongVA-7B-DPO | 49 | #2188 | |
| 8 | Kimi K2.5 | Thinking | 48.5 | #266 |
| 9 | Qwen 3 VL 8B Instruct | 47.5 | #1029 | |
| 10 | Qwen 3 Omni 30B A3B (Thinking) | 47.5 | ||
| 11 | Gemini 3 Flash (Preview) | 47 | #152 | |
| 12 | LLaVA-OneVision-7B | 47 | #2981 | |
| 13 | Qwen 3 VL 8B (Thinking) | 43 | #910 | |
| 14 | InternVL2.5-8B | 43 | #2370 | |
| 15 | LLaVA-Video-7B-Qwen2 | 42.5 | #1978 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=egointrospect-memory-recall-prediction · How It Works · Data refreshed daily, snapshot 2026-11-21.