ActivityNet-QA
Gemini 2.5 Pro Jun 2025 66.7% leads since Jul 2025 Dropped Models answer open-ended questions about web videos of human activities, from 58,000 QA pairs on 5,800 videos. Scored by accuracy.
Many recent papers grade answers with a GPT judge, so scores vary by protocol.
In the authors’ words
Recent developments in modeling language and vision have been successfully applied to image question answering. It is both crucial and natural to extend this research direction to the video domain for video question answering (VideoQA). Compared to the image domain where large scale and fully annotated benchmark datasets exists, VideoQA datasets are limited to small scale and are automatically generated, etc. These limitations restrict their applicability in practice. Here we introduce ActivityNet-QA, a fully annotated and large scale VideoQA dataset. The dataset consists of 58,000 QA pairs on 5,800 complex web videos derived from the popular ActivityNet dataset. We present a statistical analysis of our ActivityNet-QA dataset and conduct extensive experiments on it by comparing existing VideoQA baselines. Moreover, we explore various video representation strategies to improve VideoQA performance, especially for long videos. The dataset is available at https://github.com/MILVLG/activitynet-qa
- Level
- Everyday Most adults, without training
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Short answer A number, word or short string matched to a key.
Gemini 2.5 Pro leads with 66.7% among models released by Jun 2025
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
The newest model with a score here came out in Jun 2025, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6 Astra
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- and 16 more
- 1 Gemini 2.5 Pro Jun 2025 66.7%
- 2 Gemini 2.5 Flash Jun 2025 65.1%
- 3 GPT-4o May 2024 61.9%
- 4 Gemini 1.5 Pro May 2024 57.5%
- 5 Gemini 1.5 Pro Feb 2024 56.7%
- 6 Gemma 3 27B 52.8%
- 7 Gemma 3 12B 50.4%
- 8 Gemma 3 4B 46.3%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Gemini 2.5 Pro Jun 2025 | 66.7% | Jul 2025 | Model card | |
| 2 | Gemini 2.5 Flash Jun 2025 | 65.1% | Jul 2025 | Model card | |
| 3 | GPT-4o May 2024 | 61.9% | May 2024 | Model card | |
| 4 | Gemini 1.5 Pro May 2024 | 57.5% | May 2024 | Model card | |
| 5 | Gemini 1.5 Pro Feb 2024 | 56.7% | Feb 2024 | Model card | |
| 6 | Gemma 3 27B | 52.8% | Mar 2025 | Model card | |
| 7 | Gemma 3 12B | 50.4% | Mar 2025 | Model card | |
| 8 | Gemma 3 4B | 46.3% | Mar 2025 | Model card |
3 models have held number one since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No launch post has listed it since May 2024
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 14 May 2024 | Gemini 1.5 Pro | 57.5 | ActivityNet-QA (test), 0-shot |
| 13 May 2024 | GPT-4o | 61.9 | ActivityNet (%) (test) |
| 15 Feb 2024 | Gemini 1.5 Pro | 56.7 | ActivityNet-QA (test), 0-shot |
5 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 7 Jul 2025 | Gemini 2.5 Pro technical report | 66.7 | ActivityNet-QA, test, 0-shot, visual-only (Table 6) | Capabilities |
| 7 Jul 2025 | Gemini 2.5 Flash technical report | 65.1 | ActivityNet-QA, test, 0-shot, visual-only (Table 6) | Capabilities |
| 25 Mar 2025 | Gemma 3 4B technical report | 46.3 | ActivityNet-QA, IT, 0-shot, 16 frames, GPT evaluation (Table 17) | Capabilities |
| 25 Mar 2025 | Gemma 3 27B technical report | 52.8 | ActivityNet-QA, IT, 0-shot, 16 frames, GPT evaluation (Table 17) | Capabilities |
| 25 Mar 2025 | Gemma 3 12B technical report | 50.4 | ActivityNet-QA, IT, 0-shot, 16 frames, GPT evaluation (Table 17) | Capabilities |
More in Audio and video
- Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy.
Qwen 3.8 Max 88.7% - LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization. Gemini 4 Argon 91.7%
- MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy. Qwen 3.8 Max 82.4%
- MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy. Qwen 3.8 Max 90.8%
- VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules. Qwen3.5-Omni-Plus 93.1%
- TVBench In use Models answer multiple-choice questions on 10 temporal video tasks, such as action counting and ordering, that static frames cannot solve. Scored by accuracy. Qwen 3.8 Max 81.9%