Benchmark Almanac

1H-VideoQA

Gemini 3.7 Flash 89.1% leads since Aug 2026 Only in system cards

Models answer questions about videos about one hour long, so the answer needs evidence from far parts of the video.

Google DeepMind released the 101 questions as a Kaggle benchmark in 2026-06. Earlier Gemini reports used it as an in-house set.

Made by Google DeepMind Released Feb 2024 Long videos All tasks public Last in a system card Jul 2025 Last model added Oct 2026 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Multiple choice
google-deepmind/1h-videoqa 1 stars Paper Dataset Leaderboard
20
Models scored on its own leaderboard
2
System cards report it, from 1 lab

Gemini 3.7 Flash leads with 89.1%

Best score per model from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best.

20 models. Scroll sideways for the rest.

  1. 89.1 Gemini 3.7 Flash
  2. 87.1 Gemini 3.8 Flash
  3. 87.0 Gemini 4 Argon
  4. 82.2 Gemini 3.6 Flash
  5. 80.2 Gemini 3.5 Flash
  6. 79.2 Gemini 3 Flash
  7. 79.0 Gemini 2.5 Pro
  8. 78.2 Gemini 3.1 Pro
  9. 78.2 Gemini 3.1 Flash-Lite
  10. 75.3 Gemini 3.5 Flash-Lite
  11. 75.3 GPT-6 Astra
  12. 71.3 Claude Opus 4.8
  13. 68.3 Claude Haiku 5.5
  14. 67.3 Claude Sonnet 4.6
  15. 67.3 GPT-5.6 Terra
  16. 66.3 GPT-6 Sol
  17. 62.4 GPT-5.6 Sol
  18. 60.0 Grok 4.6
  19. 59.4 Grok 4.7
  20. 57.4 GPT-6 Luna
  1. 1 Gemini 3.7 Flash 89.1%
  2. 2 Gemini 3.8 Flash 87.1%
  3. 3 Gemini 4 Argon 87%
  4. 4 Gemini 3.6 Flash 82.2%
  5. 5 Gemini 3.5 Flash 80.2%
  6. 6 Gemini 3 Flash 79.2%
  7. 7 Gemini 2.5 Pro Jun 2025 79%
  8. 8 Gemini 3.1 Pro 78.2%
  9. 9 Gemini 3.1 Flash-Lite 78.2%
  10. 10 Gemini 3.5 Flash-Lite 75.3%
  11. 11 GPT-6 Astra 75.3%
  12. 12 Claude Opus 4.8 71.3%
  13. 13 Claude Haiku 5.5 68.3%
  14. 14 Claude Sonnet 4.6 67.3%
  15. 15 GPT-5.6 Terra 67.3%
  16. 16 GPT-6 Sol 66.3%
  17. 17 GPT-5.6 Sol 62.4%
  18. 18 Grok 4.6 60%
  19. 19 Grok 4.7 59.4%
  20. 20 GPT-6 Luna 57.4%
#ModelScoreSettingDateSource
1 Gemini 3.7 Flash 89.1% Aug 2026 Official leaderboard
2 Gemini 3.8 Flash 87.1% Sep 2026 Official leaderboard
3 Gemini 4 Argon 87% Sep 2026 Official leaderboard
4 Gemini 3.6 Flash 82.2% Jul 2026 Official leaderboard
5 Gemini 3.5 Flash 80.2% May 2026 Official leaderboard
6 Gemini 3 Flash 79.2% Dec 2025 Official leaderboard
7 Gemini 2.5 Pro Jun 2025 79% Jun 2025 Official leaderboard
8 Gemini 3.1 Pro 78.2% Feb 2026 Official leaderboard
9 Gemini 3.1 Flash-Lite 78.2% Mar 2026 Official leaderboard
10 Gemini 3.5 Flash-Lite 75.3% Jul 2026 Official leaderboard
11 GPT-6 Astra 75.3% Sep 2026 Official leaderboard
12 Claude Opus 4.8 71.3% May 2026 Official leaderboard
13 Claude Haiku 5.5 68.3% Oct 2026 Official leaderboard
14 Claude Sonnet 4.6 67.3% Feb 2026 Official leaderboard
15 GPT-5.6 Terra 67.3% Jul 2026 Official leaderboard
16 GPT-6 Sol 66.3% Sep 2026 Official leaderboard
17 GPT-5.6 Sol 62.4% Jul 2026 Official leaderboard
18 Grok 4.6 60% Aug 2026 Official leaderboard
19 Grok 4.7 59.4% Sep 2026 Official leaderboard
20 GPT-6 Luna 57.4% Sep 2026 Official leaderboard

Labs report scores for 2 models in their cards. We plot every source in the charts below.

5 models have held number one since its release in Feb 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind
Own leaderboard Passed 85% in Aug 2026, by Gemini 3.7 Flash
Gemini 3.7 Flash 89.1%

The best score passed 85% in Aug 2026

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3.7 Flash Gemini 3.6 Flash Gemini 2.5 Pro 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3.7 Flash Gemini 3.6 Flash

Labs report it only in system cards

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.

2 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.

DateDocumentScoreAs the document writes itSection
7 Jul 2025 Gemini 2.5 Pro technical report 81 1H-VideoQA, test, 0-shot (Table 6) Capabilities
7 Jul 2025 Gemini 2.5 Flash technical report 67.5 1H-VideoQA, test, 0-shot (Table 6) Capabilities

More in Audio and video

All 168 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%