Benchmark Almanac

ViMU

o4-mini 46.9% leads since May 2026
Not in model cards

Models explain what a short video implies and select the cues that support their reading. They also identify rhetorical devices and social signals.

The authors report four task scores for 16 models on 588 videos. A model judge grades explanations. The other tasks use answer keys.

Released
May 2026
Kind
Video understanding
Level
Everyday
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Video → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
In views
Vision
Strongest model tested
GPT-5.2 #51 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
16
Models scored on its own leaderboard

o4-mini leads with 46.9% among models released by March 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in March 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 o4-mini 46.9%
  2. 2 Grok 4.1 Fast 46.3%
  3. 3 Qwen3.5-27B 45.9%
  4. 4 GPT-5.2 44.7%
  5. 5 Gemini 3 Flash 44.3%
  6. 6 Qwen3-VL-32B-Instruct 41.6%
  7. 7 Seed2.0 Lite (0215) 40.6%
  8. 8 MiMo-V2-Omni 38.1%
  9. 9 GPT-5.4 Mini 36.6%
  10. 10 Gemma 3 27B 36.4%
  11. 11 Ministral 3 14B 35.5%
  12. 12 Ministral 8B 34.8%
  13. 13 GLM-4.5V 25.9%
  14. 14 Gemma 3 4B 23.3%
  15. 15 Claude 3 Haiku 22.9%
  16. 16 GPT-4.1 nano 20.9%

Source: Official ViMU leaderboard

#ModelScoreSettingDateSource
1 o4-mini 46.9% Apr 2025 Official leaderboard
2 Grok 4.1 Fast 46.3% Nov 2025 Official leaderboard
3 Qwen3.5-27B 45.9% Feb 2026 Official leaderboard
4 GPT-5.2 44.7% Dec 2025 Official leaderboard
5 Gemini 3 Flash 44.3% Dec 2025 Official leaderboard
6 Qwen3-VL-32B-Instruct 41.6% Oct 2025 Official leaderboard
7 Seed2.0 Lite (0215) 40.6% Feb 2026 Official leaderboard
8 MiMo-V2-Omni 38.1% Mar 2026 Official leaderboard
9 GPT-5.4 Mini 36.6% Mar 2026 Official leaderboard
10 Gemma 3 27B 36.4% Mar 2025 Official leaderboard
11 Ministral 3 14B 35.5% Dec 2025 Official leaderboard
12 Ministral 8B 34.8% Oct 2024 Official leaderboard
13 GLM-4.5V 25.9% Aug 2025 Official leaderboard
14 Gemma 3 4B 23.3% Mar 2025 Official leaderboard
15 Claude 3 Haiku 22.9% Mar 2024 Official leaderboard
16 GPT-4.1 nano 20.9% Apr 2025 Official leaderboard

Rankings by part

We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as OE, EG and RM. GPT-5.2 leads 2 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

OE

16 models we track have a score here.

  1. 1 GPT-5.2 ↑3 73.2%
  2. 2 GPT-5.4 Mini ↑7 66.2%
  3. 3 o4-mini ↓2 65.3%
  4. 4 Qwen3-VL-32B-Instruct ↑2 64.1%
  5. 5 MiMo-V2-Omni ↑3 64.1%
  6. 6 Qwen3.5-27B ↓3 62.8%
  7. 7 Gemini 3 Flash ↓2 62.5%
  8. 8 GLM-4.5V ↑5 62.5%
  9. 9 Seed2.0 Lite (0215) ↓2 60.8%
  10. 10 Grok 4.1 Fast ↓8 57.6%

Number one over time

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

OpenAI
Own leaderboard Not tested since May 2026
o4-mini 46.9%

Source: Official ViMU leaderboard

Scores over time

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

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) o4-mini Gemma 3 27B Ministral 8B Claude 3 Haiku 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) o4-mini Gemma 3 27B

Source: Official ViMU leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Audio and video

All 695 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%
  • FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better. Phi-4-multimodal 4.00%
  • 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%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 8 months ago, in March 2026, for MiMo-V2-Omni. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (16 entries).
  • Pass Models scored 16 models we track.
  • Warning Newest score The newest came 8 months ago, in March 2026, for MiMo-V2-Omni.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.2 (#51).
  • Pass Headroom The best is 46.9%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on liqiiiii.github.io: 16 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check