OmniBehavior - Live Streaming Binary Behaviors: leaderboard
Metric: F1 (%) on binary engagement decisions (such as comment) in the live-streaming scenario, for predicting the real behaviour of 200 Kuaishou users (three months of interleaved video, live-stream, ads and e-commerce traces, 8,143 actions per user on average) from the user profile, history and current scenario, 6,000 balanced tasks, 32k context window, temperature 0.1; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.5 | 31.7 |
| 2 | GLM-4.7 | 28.97 |
| 3 | Claude Sonnet 4 | 28.86 |
| 4 | GPT-5.2 | 28.63 |
| 5 | Kimi K2 0905 | 28.6 |
| 6 | GPT-4o | 28.15 |
| 7 | Claude Haiku 4.5 | 26.11 |
| 8 | Gemini 3 Flash | 25.61 |
| 9 | Claude Sonnet 4.5 | 25 |
| 10 | Qwen 3 235B A22B 2507 Instruct | 23.84 |
Interactive version: theaggregate.ai/benchmark?slug=omnibehavior-live-streaming-binary-behaviors · How It Works · Data refreshed daily, snapshot 2026-10-07.