ExtremeWhenBench

What is ExtremeWhenBench?

ExtremeWhenBench evaluates natural-language temporal grounding in hour-long videos, with 2,273 open-form queries over 194 videos. Metrics include mIoU and R@k…

Released

2026-06-10

Evaluates

General AI, Vision & 3D, cs.CV

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.12300

ExtremeWhenBench paper and code

• Paper https://arxiv.org/abs/2606.12300
• Code https://github.com/naver-ai/ExtremeWhenBench

Which sources cite ExtremeWhenBench?

• ExtremeWhenBench

Related General AI, Vision & 3D, cs.CV benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 360CityArena
• 3DCodeBench
• 5-Dialects-BN
• A3S-Bench
• AARRI-Bench
• ActBench
• ActiveSciBench