Benchmark Almanac

DesignArena Video Editing

MiniMax H3 1,373
Not in model cards

Video models edit a user video following a text instruction. Users compare the two results and vote for the better one.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Video generation
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Scores refreshed
November 12th, 2026
20
Entries on its own leaderboard, 5 models we track

MiniMax H3 leads with 1,373

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

20 entries. Scroll sideways for the rest. Bars start at 900.

  1. 1,373 MiniMax H3
  2. 1,335 Gemini Omni Flash
  3. 1,327 Gemini Omni 1.1 Flash
  4. 1,270 FLUX 3 Video Edit
  5. 1,267 Wan 2.7
  6. 1,263 HappyHorse 1.0
  7. 1,226 Aleph 2.0
  8. 1,194 Grok Imagine
  9. 1,188 Kling O3
  10. 1,183 Ray 3.14
  11. 1,155 LTX-2 19B
  12. 1,152 Kling O1 Edit
  13. 1,150 p-video-edit
  14. 1,133 Gen-4 Aleph
  15. 1,103 Lucy 2.5
  16. 1,074 Wan Vace
  17. 1,073 Lucy Edit (Pro)
  18. 1,007 Kling v2.6 Pro
  19. 991 Kling v2.5 Turbo Pro
  20. 903 Krea Wan 14b

Bars start at 900.

  1. 1 MiniMax H3 max 1,373
  2. 2 Gemini Omni Flash 1,335
  3. 3 Gemini Omni 1.1 Flash 1,327
  4. 4 FLUX 3 Video Edit 1,270
  5. 5 Wan 2.7 1,267
  6. 6 HappyHorse 1.0 1,263
  7. 7 Aleph 2.0 1,226
  8. 8 Grok Imagine 1,194
  9. 9 Kling O3 1,188
  10. 10 Ray 3.14 1,183
  11. 11 LTX-2 19B 1,155
  12. 12 Kling O1 Edit 1,152
  13. 13 p-video-edit draft 1,150
  14. 14 Gen-4 Aleph 1,133
  15. 15 Lucy 2.5 1,103
  16. 16 Wan Vace 1,074
  17. 17 Lucy Edit (Pro) 1,073
  18. 18 Kling v2.6 Pro 1,007
  19. 19 Kling v2.5 Turbo Pro 991
  20. 20 Krea Wan 14b 903

Source: Official DesignArena Video Editing leaderboard

#EntryScoreSettingDateSource
1 MiniMax H3 1,373 max Official leaderboard
2 Gemini Omni Flash 1,335 Official leaderboard
3 Gemini Omni 1.1 Flash 1,327 Official leaderboard
4 FLUX 3 Video Edit 1,270 Official leaderboard
5 Wan 2.7 1,267 Official leaderboard
6 HappyHorse 1.0 1,263 Official leaderboard
7 Aleph 2.0 1,226 Official leaderboard
8 Grok Imagine 1,194 Official leaderboard
9 Kling O3 1,188 Official leaderboard
10 Ray 3.14 1,183 Official leaderboard
11 LTX-2 19B 1,155 Official leaderboard
12 Kling O1 Edit 1,152 Official leaderboard
13 p-video-edit 1,150 draft Official leaderboard
14 Gen-4 Aleph 1,133 Official leaderboard
15 Lucy 2.5 1,103 Official leaderboard
16 Wan Vace 1,074 Official leaderboard
17 Lucy Edit (Pro) 1,073 Official leaderboard
18 Kling v2.6 Pro 1,007 Official leaderboard
19 Kling v2.5 Turbo Pro 991 Official leaderboard
20 Krea Wan 14b 903 Official leaderboard

Scores over time

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

No dated values yet.

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Audio and video

All 232 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%