OVBench
OVBench is a benchmark tailored for real-time video understanding:
How to run
ltb · ovbench
$ npm i -g lettokenburn-cli$ ltb bench info ovbench
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
6 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| Seed1.5-VL thinking | 60.0 | N/A | Independent | medium | May 11, 2025 | arxiv.org |
| Gemini 1.5 Flash | 50.7 | N/A | Independent | medium | Mar 8, 2024 | arxiv.org |
| Qwen2-VL | 49.7 | N/A | Independent | medium | Sep 18, 2024 | arxiv.org |
| LLaVA-OneVision | 49.5 | N/A | Independent | medium | Aug 6, 2024 | arxiv.org |
| InternVL2 | 48.7 | N/A | Independent | medium | Dec 6, 2024 | arxiv.org |
| LLaMA-VID | 41.9 | N/A | Independent | medium | Nov 28, 2023 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsLoading community runs…
About
OVBench is a benchmark tailored for real-time video understanding:
- Memory, Perception, and Prediction of Temporal Contexts: Questions are framed to reference the present state of entities, requiring models to memorize/perceive/predict past/present/future temporal contexts over time.
- Dynamic Spatio-temporal Interaction: The benchmark demands precise real-time interactions with video content, where actions, objects, and events must be understood in the context of their spatial and temporal relationships.
- Contextual Awareness at Specific Moments: Real-time questions are contextual, changing based on the specific timestamp they are asked, requiring a deep understanding of how temporal context evolves.