S-EMBER - Grounded QA: leaderboard

Streaming egocentric memory retrieval (S-EMBER, Wang et al., Meta, 2026): 9,448 questions about 388 hours of first-person video from Ray-Ban Meta glasses, each asked at a trigger time and answerable only from what the wearer saw earlier, with open models given 128 uniformly sampled frames and proprietary models their platform limits. A question counts only when the free-form answer is correct and the predicted evidence interval overlaps the annotated one with IoU of at least 0.5; share of questions from 0 to 1, higher is better.

Metric: Grounded QA at IoU 0.5 (0-1; share of questions answered correctly from an interval with IoU of at least 0.5) (S-EMBER: 9,448 questions on 3,141 egocentric Ray-Ban Meta videos, answered from the stream up to the question's trigger time; open models at 128 uniformly sampled frames, proprietary models at their platform limits). Source: arxiv.org. Saturation forecast: Around January 2028. 5 models tracked.

Top models

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=s-ember-grounded-qa · How It Works · Data refreshed daily, snapshot 2026-11-12.