Benchmark Almanac

MOSEv2

SAM2Long-L 51.5% tops its leaderboard
Not in model cards
Version 2

Models track and segment objects in videos with complex scenes. Objects may disappear, reappear, or blend into a crowd.

Made by
Henghui Ding and colleagues
Released
August 2025
Kind
Video understanding
Level
Everyday
Scored
Checked automatically
Answers
Prediction
Tasks
All public
Measures
Capability
In and out
Video → Images
Interaction
Single turn
Run by
Anyone
Based on
MOSE
20
Entries in its paper, 0 models we track

SAM2Long-L tops the leaderboard with 51.5%

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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.

  1. 1 SAM2Long-L Mask initialization 51.5%
  2. 2 DAM4SAM-L Mask initialization 51.2%
  3. 3 SAMURAI-L Mask initialization 51.1%
  4. 4 SAM2-L Mask initialization 49.7%
  5. 5 SAM2Long-B+ Mask initialization 48.6%
  6. 6 DAM4SAM-B+ Mask initialization 47.9%
  7. 7 SAMURAI-B+ Mask initialization 47.4%
  8. 8 SAM2-B+ Mask initialization 46%
  9. 9 Cutie-B Mask initialization 42.8%
  10. 10 Cutie+SAM Box initialization 42.3%
  11. 11 DEVA Mask initialization 38.3%
  12. 12 JointFormer Mask initialization 37.7%
  13. 13 XMem Mask initialization 36.3%
  14. 14 XMem++ Mask initialization 34.2%
  15. 15 DeAOT-L Mask initialization 32.6%
  16. 16 RDE Mask initialization 32%
  17. 17 AOT-L Mask initialization 30.2%
  18. 18 STCN Mask initialization 29.7%
  19. 19 EntitySAM Unsupervised 28.2%
  20. 20 UniVS Box initialization 16.4%

Source: The MOSEv2 paper

#EntryScoreSettingDateSource
1 SAM2Long-L 51.5% Mask initialization Official leaderboard
2 DAM4SAM-L 51.2% Mask initialization Official leaderboard
3 SAMURAI-L 51.1% Mask initialization Official leaderboard
4 SAM2-L 49.7% Mask initialization Official leaderboard
5 SAM2Long-B+ 48.6% Mask initialization Official leaderboard
6 DAM4SAM-B+ 47.9% Mask initialization Official leaderboard
7 SAMURAI-B+ 47.4% Mask initialization Official leaderboard
8 SAM2-B+ 46% Mask initialization Official leaderboard
9 Cutie-B 42.8% Mask initialization Official leaderboard
10 Cutie+SAM 42.3% Box initialization Official leaderboard
11 DEVA 38.3% Mask initialization Official leaderboard
12 JointFormer 37.7% Mask initialization Official leaderboard
13 XMem 36.3% Mask initialization Official leaderboard
14 XMem++ 34.2% Mask initialization Official leaderboard
15 DeAOT-L 32.6% Mask initialization Official leaderboard
16 RDE 32% Mask initialization Official leaderboard
17 AOT-L 30.2% Mask initialization Official leaderboard
18 STCN 29.7% Mask initialization Official leaderboard
19 EntitySAM 28.2% Unsupervised Official leaderboard
20 UniVS 16.4% Box initialization Official 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.

Versions

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

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (20 entries).
  • Pass Models scored 20 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom No scores yet.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 20 entries at the last read, none a model we track.
  • 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