EgoSchema
video
+
+
+
+
About
EgoSchema is a diagnostic benchmark for very long-form video understanding that evaluates AI models' ability to comprehend extended video content and answer complex questions about temporal relationships, causality, and narrative structure. This benchmark tests multimodal reasoning over long video sequences, measuring how well models can analyze sustained visual narratives and extract meaningful insights from extended egocentric video content.
+
+
+
+
Evaluation Stats
Total Models9
Organizations4
Verified Results0
Self-Reported9
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
9 models
Top Score
77.9%
Average Score
70.3%
High Performers (80%+)
0Top Organizations
#1Alibaba Cloud / Qwen Team
3 models
74.2%
#2OpenAI
1 model
72.2%
#3Amazon
2 models
71.8%
#4Google
3 models
64.8%
+
+
+
+
Leaderboard
9 models ranked by performance on EgoSchema
| License | Links | ||||
|---|---|---|---|---|---|
| Aug 29, 2024 | tongyi-qianwen | 77.9% | |||
| Jan 26, 2025 | tongyi-qianwen | 76.2% | |||
| Aug 6, 2024 | Proprietary | 72.2% | |||
| Nov 20, 2024 | Proprietary | 72.1% | |||
| Dec 1, 2024 | Proprietary | 71.5% | |||
| Nov 20, 2024 | Proprietary | 71.4% | |||
| Mar 27, 2025 | Apache 2.0 | 68.6% | |||
| Feb 5, 2025 | Proprietary | 67.2% | |||
| Feb 15, 2024 | Proprietary | 55.7% |
