Benchmark Almanac

EvaOmni

Gemini 3.7 Flash 79.4% leads since October 2026
Not in model cards

Models answer questions about videos and identify the audio and visual evidence behind their answers. The score accounts for both correctness and evidence timing.

Made by
Xiaoxing Hu and colleagues
Released
October 2026
Kind
Audio understanding and voice
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Video, Audio → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Gemini 4 Argon #10 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
23
Entries on its own leaderboard, 22 models we track

Gemini 3.7 Flash leads with 79.4% among models released by September 2026 Gemini 3.7 Flash leads with 79.4% among models released by September 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 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. This view shows only models we track, each at its best.

The newest model with a score here came out in September 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 Gemini 3.7 Flash 79.4%
  2. 2 Gemini 3.8 Flash 78%
  3. 3 Gemini 3.6 Flash 74.4%
  4. 4 Gemini 3.1 Pro 72.3%
  5. 5 Gemini 2.5 Pro Jun 2025 71%
  6. 6 Doubao-Seed-2.1-Lite 70.9%
  7. 7 Gemini 3.5 Flash 68.5%
  8. 8 Gemini 4 Argon 61.4%
  9. 9 Qwen3.8-Omni-Flash 58.2%
  10. 10 MiMo-V2.5 56.8%
  11. 11 Doubao Seed 2.0 Lite Apr 2026 55.4%
  12. 12 MiMo-V2.6 Flash 53.6%
  13. 13 MiMo-V2.6 Pro 51.8%
  14. 14 Qwen3.5-Omni-Plus 49%
  15. 15 Gemini 3.5 Flash-Lite 48.3%
  16. 16 Gemini 2.5 Flash Sep 2025 48.1%
  17. 17 Gemini 3.1 Flash-Lite 47.8%
  18. 18 Gemini 2.5 Flash-Lite Sep 2025 39.7%
  19. 19 Qwen3.5-Omni-Flash 37.7%
  20. 20 Qwen3-Omni-30B-A3B 25.2%
  21. 21 Qwen3-Omni-30B-A3B-Thinking 24.8%
  22. 22 Nemotron 3 Nano Omni 30B-A3B 24.2%
  23. 23 Qwen2.5-Omni-7B 16.4%

Source: Official EvaOmni leaderboard

#EntryScoreSettingDateSource
1 Gemini 3.7 Flash 79.4% Aug 2026 Official leaderboard
2 Gemini 3.8 Flash 78% Sep 2026 Official leaderboard
3 Gemini 3.6 Flash 74.4% Jul 2026 Official leaderboard
4 Gemini 3.1 Pro 72.3% Feb 2026 Official leaderboard
5 Gemini 2.5 Pro Jun 2025 71% Jun 2025 Official leaderboard
6 Doubao-Seed-2.1-Lite closed weights 70.9% Official leaderboard
7 Gemini 3.5 Flash 68.5% May 2026 Official leaderboard
8 Gemini 4 Argon 61.4% Sep 2026 Official leaderboard
9 Qwen3.8-Omni-Flash 58.2% Sep 2026 Official leaderboard
10 MiMo-V2.5 56.8% Apr 2026 Official leaderboard
11 Doubao Seed 2.0 Lite Apr 2026 55.4% Apr 2026 Official leaderboard
12 MiMo-V2.6 Flash 53.6% Sep 2026 Official leaderboard
13 MiMo-V2.6 Pro 51.8% Sep 2026 Official leaderboard
14 Qwen3.5-Omni-Plus 49% Mar 2026 Official leaderboard
15 Gemini 3.5 Flash-Lite 48.3% Jul 2026 Official leaderboard
16 Gemini 2.5 Flash Sep 2025 48.1% Sep 2025 Official leaderboard
17 Gemini 3.1 Flash-Lite 47.8% Mar 2026 Official leaderboard
18 Gemini 2.5 Flash-Lite Sep 2025 39.7% Sep 2025 Official leaderboard
19 Qwen3.5-Omni-Flash 37.7% Mar 2026 Official leaderboard
20 Qwen3-Omni-30B-A3B 25.2% Sep 2025 Official leaderboard
21 Qwen3-Omni-30B-A3B-Thinking 24.8% Sep 2025 Official leaderboard
22 Nemotron 3 Nano Omni 30B-A3B 24.2% Apr 2026 Official leaderboard
23 Qwen2.5-Omni-7B 16.4% Mar 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 3.7 Flash 79.4% Aug 2026 Official leaderboard
2 Gemini 3.8 Flash 78% Sep 2026 Official leaderboard
3 Gemini 3.6 Flash 74.4% Jul 2026 Official leaderboard
4 Gemini 3.1 Pro 72.3% Feb 2026 Official leaderboard
5 Gemini 2.5 Pro Jun 2025 71% Jun 2025 Official leaderboard
6 Gemini 3.5 Flash 68.5% May 2026 Official leaderboard
7 Gemini 4 Argon 61.4% Sep 2026 Official leaderboard
8 Qwen3.8-Omni-Flash 58.2% Sep 2026 Official leaderboard
9 MiMo-V2.5 56.8% Apr 2026 Official leaderboard
10 Doubao Seed 2.0 Lite Apr 2026 55.4% Apr 2026 Official leaderboard
11 MiMo-V2.6 Flash 53.6% Sep 2026 Official leaderboard
12 MiMo-V2.6 Pro 51.8% Sep 2026 Official leaderboard
13 Qwen3.5-Omni-Plus 49% Mar 2026 Official leaderboard
14 Gemini 3.5 Flash-Lite 48.3% Jul 2026 Official leaderboard
15 Gemini 2.5 Flash Sep 2025 48.1% Sep 2025 Official leaderboard
16 Gemini 3.1 Flash-Lite 47.8% Mar 2026 Official leaderboard
17 Gemini 2.5 Flash-Lite Sep 2025 39.7% Sep 2025 Official leaderboard
18 Qwen3.5-Omni-Flash 37.7% Mar 2026 Official leaderboard
19 Qwen3-Omni-30B-A3B 25.2% Sep 2025 Official leaderboard
20 Qwen3-Omni-30B-A3B-Thinking 24.8% Sep 2025 Official leaderboard
21 Nemotron 3 Nano Omni 30B-A3B 24.2% Apr 2026 Official leaderboard
22 Qwen2.5-Omni-7B 16.4% Mar 2025 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: Level 1 · Retrieval & grounding, Level 2 · Basic understanding and Level 3 · Complex reasoning. Each ranks models on one part of the test. Gemini 3.7 Flash leads 2 of them, and Gemini 3.8 Flash leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Level 1 · Retrieval & grounding

22 models we track have a score here.

  1. 1 Gemini 3.7 Flash 85.3%
  2. 2 Gemini 3.8 Flash 83.4%
  3. 3 Gemini 3.6 Flash 80.9%
  4. 4 Gemini 2.5 Pro Jun 2025 ↑1 76.5%
  5. 5 Gemini 3.1 Pro ↓1 74.7%
  6. 6 Gemini 4 Argon ↑1 67.1%
  7. 7 MiMo-V2.5 ↑2 65.4%
  8. 8 Qwen3.8-Omni-Flash 65%
  9. 9 Gemini 3.5 Flash ↓3 64.1%
  10. 10 MiMo-V2.6 Flash ↑1 61.3%

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 Not tested since Oct 2026

Source: Official EvaOmni 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% 2025 2026 Near the ceiling (85%) Gemini 3.7 Flash Gemini 3.6 Flash Gemini 3.1 Pro Gemini 2.5 Pro Qwen2.5-Omni-7B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3.7 Flash Gemini 3.1 Pro

Source: Official EvaOmni 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

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (23 entries).
  • Pass Models scored 22 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from September 2026, for Gemini 4 Argon.
  • Note Frontier models Only 1 of the top 10 has a score. The strongest tested is Gemini 4 Argon (#10).
  • Pass Headroom The best is 79.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on xiaoxing2001.github.io: 23 rows for models we track and 1 other entry 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