VideoJev Evaluation - ActivityNet Captions TimeLens - mIoU: leaderboard
VideoJev Evaluation - ActivityNet Captions TimeLens - mIoU. Refined TimeLens annotations; generative models Instruct mode; VideoJev Joint inference. Source Table1; latency is not capability. Metric: mIoU (%). Primary published table; source configurations remain distinct.
Metric: mIoU (%). Source: arxiv.org. Saturation forecast: Around 2031. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 58.1 | #150 |
| 2 | Qwen 3.6 35B A3B | 56 | #217 |
| 3 | Qwen 3.5 4B | 55.6 | #463 |
| 4 | Qwen 3.5 35B A3B | 54.4 | #250 |
| 5 | GPT-5 | 42.9 | #94 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=videojev-evaluation-activitynet-captions-timelens-miou · How It Works · Data refreshed daily, snapshot 2026-10-16.