Benchmark Almanac

LMArena Text-to-Video Arena

Gemini Omni 1.1 Flash 1,516 leads since Sep 2026 Not in model cards

Video generation models make clips from the same text prompt, anonymously and side by side. Votes from logged-in users give Elo-style ratings.

Video Arena moved to the web on 2026-01-21 with 15 video models. It ran earlier in another venue, per the blog.

Made by Arena (formerly LMArena) Released Jan 2026 Video generation All tasks public Last model added Sep 2026 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Ranked by votes People pick the better of two answers, as in an arena, and the votes are turned into ratings.
Answers
Open-ended Free-form text, documents, designs or media judged as a whole.
46
Entries on its own leaderboard, 45 models we track

Gemini Omni 1.1 Flash leads with 1,516

Best score per entry from the benchmark's own leaderboard, read on 2 Nov 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.

Bars start at 1,000.

  1. 1 Gemini Omni 1.1 Flash 1,516
  2. 2 Gemini Omni Flash 1,513
  3. 3 Flux 3 Video 1,493
  4. 4 Grok Imagine Video 1.5 Agent 1,492
  5. 5 Dreamina Seedance 2.0 1,479
  6. 6 Wan 3.0 1,476
  7. 7 Dreamina Seedance 2.5 1,474
  8. 8 MiniMax H3 1,460
  9. 9 Muse Video 1,456
  10. 10 HappyHorse 1.0 1,427
  11. 11 Sora 2 Pro 1,368
  12. 12 Veo 3.1 Audio 1,364
  13. 13 Veo 3.1 Fast Audio 1,362
  14. 14 Veo 3 Fast Audio 1,348
  15. 15 Sora 2 1,343
  16. 16 Grok Imagine Video 1,342
  17. 17 Veo 3 Audio 1,340
  18. 18 Wan 2.7 T2V 1,337
  19. 19 Wan 2.6 T2V 1,327
  20. 20 Seedance v1.5 Pro 1,256
  21. 21 Veo 3 1,254
  22. 22 Veo 3 Fast 1,248
  23. 23 Wan 2.5 T2V Preview 1,245
  24. 24 PixVerse v5.6 1,240
  25. 25 Runway Gen-4.5 1,225
  26. 26 Kling 2.5 Turbo 1,220
  27. 27 Kling 2.6 Pro 1,216
  28. 28 p-video 1,207
  29. 29 Ray 3 1,206
  30. 30 Hailuo 2.3 1,206
  31. 31 Kling O1 Pro 1,205
  32. 32 Hailuo 02 Pro 1,198
  33. 33 Seedance v1 Pro 1,191
  34. 34 Hailuo 02 Standard 1,181
  35. 35 Kandinsky 5.0 T2V Pro 1,175
  36. 36 HunyuanVideo 1.5 1,169
  37. 37 Veo 2 1,164
  38. 38 Kling v2.1 Master 1,163
  39. 39 LTX-2 19B 1,154
  40. 40 Wan 2.2 A14B 1,132
  41. 41 Seedance v1 Lite 1,113
  42. 42 Kandinsky 5.0 T2V Lite 1,113
  43. 43 Sora 1,069
  44. 44 Ray2 1,065
  45. 45 Pika v2.2 1,009
  46. 46 Mochi v1 1,007
#EntryScoreSettingDateSource
1 Gemini Omni 1.1 Flash 1,516 Sep 2026 Official leaderboard
2 Gemini Omni Flash 1,513 Sep 2026 Official leaderboard
3 Flux 3 Video 1,493 Sep 2026 Official leaderboard
4 Grok Imagine Video 1.5 Agent 1,492 Sep 2026 Official leaderboard
5 Dreamina Seedance 2.0 1,479 Sep 2026 Official leaderboard
6 Wan 3.0 1,476 Sep 2026 Official leaderboard
7 Dreamina Seedance 2.5 1,474 Sep 2026 Official leaderboard
8 MiniMax H3 1,460 Sep 2026 Official leaderboard
9 Muse Video 1,456 Sep 2026 Official leaderboard
10 HappyHorse 1.0 1,427 Sep 2026 Official leaderboard
11 Sora 2 Pro 1,368 Sep 2026 Official leaderboard
12 Veo 3.1 Audio 1,364 Sep 2026 Official leaderboard
13 Veo 3.1 Fast Audio 1,362 Sep 2026 Official leaderboard
14 Veo 3 Fast Audio 1,348 Sep 2026 Official leaderboard
15 Sora 2 1,343 Sep 2026 Official leaderboard
16 Grok Imagine Video 1,342 Sep 2026 Official leaderboard
17 Veo 3 Audio 1,340 Sep 2026 Official leaderboard
18 Wan 2.7 T2V 1,337 Sep 2026 Official leaderboard
19 Wan 2.6 T2V 1,327 Sep 2026 Official leaderboard
20 Seedance v1.5 Pro 1,256 Sep 2026 Official leaderboard
21 Veo 3 1,254 Sep 2026 Official leaderboard
22 Veo 3 Fast 1,248 Sep 2026 Official leaderboard
23 Wan 2.5 T2V Preview 1,245 Sep 2026 Official leaderboard
24 PixVerse v5.6 1,240 Sep 2026 Official leaderboard
25 Runway Gen-4.5 1,225 Sep 2026 Official leaderboard
26 Kling 2.5 Turbo 1,220 Sep 2026 Official leaderboard
27 Kling 2.6 Pro 1,216 Sep 2026 Official leaderboard
28 p-video 1,207 Sep 2026 Official leaderboard
29 Ray 3 1,206 Sep 2026 Official leaderboard
30 Hailuo 2.3 1,206 Sep 2026 Official leaderboard
31 Kling O1 Pro 1,205 Sep 2026 Official leaderboard
32 Hailuo 02 Pro 1,198 Sep 2026 Official leaderboard
33 Seedance v1 Pro 1,191 Sep 2026 Official leaderboard
34 Hailuo 02 Standard 1,181 Sep 2026 Official leaderboard
35 Kandinsky 5.0 T2V Pro 1,175 Sep 2026 Official leaderboard
36 HunyuanVideo 1.5 1,169 Sep 2026 Official leaderboard
37 Veo 2 1,164 Sep 2026 Official leaderboard
38 Kling v2.1 Master 1,163 Sep 2026 Official leaderboard
39 LTX-2 19B 1,154 Sep 2026 Official leaderboard
40 Wan 2.2 A14B 1,132 Sep 2026 Official leaderboard
41 Seedance v1 Lite 1,113 Sep 2026 Official leaderboard
42 Kandinsky 5.0 T2V Lite 1,113 Sep 2026 Official leaderboard
43 Sora 1,069 Sep 2026 Official leaderboard
44 Ray2 1,065 Sep 2026 Official leaderboard
45 Pika v2.2 1,009 Sep 2026 Official leaderboard
46 Mochi v1 1,007 Sep 2026 Official leaderboard

Number one over time

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

Google DeepMind
Own leaderboard 1 leader

Scores over time

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
1,000 1,200 1,400 1,600 1,000 1,200 1,400 1,600

No lab model card reports it yet

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

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%