ADQA-Bench
Models listen to recordings and answer 3,000 questions about speech, music, and environmental sounds. One question asks how a bird calls before it stops.
Fourteen teams submitted 36 systems. The organizers scored them by answer accuracy and published the answer key.
- Made by
- DCASE 2026 Task 5 organizers
- Released
- January 2026
- Kind
- Audio understanding and voice
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Written by crowdworkers, Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech
- Last model added
- September 2025
- Scores refreshed
- November 20th, 2026
Lim_CAU_task5_4 tops the leaderboard with 58.3% Qwen3-Omni-30B-A3B leads with 53.2%
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. The best model we track is Qwen3-Omni-30B-A3B, at 53.2%. This view shows only models we track, each at its best.
- 58.3 Lim_CAU_task5_4
- 58.1 Lim_CAU_task5_3
- 57.3 Lim_CAU_task5_1
- 57.2 Nam_IND_task5_2
- 57.1 Nam_IND_task5_4
- 57.1 Lim_CAU_task5_2
- 57.0 Hu_IOA_task5_4
- 56.7 Nam_IND_task5_3
- 56.7 Hu_IOA_task5_3
- 56.0 Yin_XJTLU_task5_1
- 55.9 Nam_IND_task5_1
- 55.8 Yin_XJTLU_task5_4
- 55.6 Yin_XJTLU_task5_2
- 55.3 Yin_XJTLU_task5_3
- 53.9 Cheng_Surrey_task5_1
- 53.5 Cheng_Surrey_task5_2
- 53.2
Qwen3-Omni-30B-A3B
- 51.6 Zhang_WHU_task5_1
- 51.1 Zhang_WHU_task5_2
- 50.6 Tathe_UIUC_task5_1
- 49.9 Hu_IOA_task5_2
- 49.8 Tathe_UIUC_task5_2
- 49.6 Hu_IOA_task5_1
- 49.6 Huang_JAIST_task5_1
- 49.1 Kim_SGU_task5_4
- 48.9 Kim_SGU_task5_3
- 48.7 Tathe_UIUC_task5_3
- 48.3 Tathe_UIUC_task5_4
- 48.0 Kim_SGU_task5_2
- 47.2 Wu_XMU_task5_1
- 46.8 Kim_SGU_task5_1
- 46.0 ZC_Inst_task5_1
- 43.4 Guan_HEU_task5_1
- 43.3 Song_BIT_task5_1
- 41.5 Song_BIT_task5_2
- 41.0 Song_BIT_task5_3
- 31.9 Xu_HUST_task5_1
- 1 Lim_CAU_task5_4 58.3%
- 2 Lim_CAU_task5_3 58.1%
- 3 Lim_CAU_task5_1 57.3%
- 4 Nam_IND_task5_2 57.2%
- 5 Nam_IND_task5_4 57.1%
- 6 Lim_CAU_task5_2 57.1%
- 7 Hu_IOA_task5_4 57%
- 8 Nam_IND_task5_3 56.7%
- 9 Hu_IOA_task5_3 56.7%
- 10 Yin_XJTLU_task5_1 56%
- 11 Nam_IND_task5_1 55.9%
- 12 Yin_XJTLU_task5_4 55.8%
- 13 Yin_XJTLU_task5_2 55.6%
- 14 Yin_XJTLU_task5_3 55.3%
- 15 Cheng_Surrey_task5_1 53.9%
- 16 Cheng_Surrey_task5_2 53.5%
- 17
Qwen3-Omni-30B-A3B 53.2%
- 18 Zhang_WHU_task5_1 51.6%
- 19 Zhang_WHU_task5_2 51.1%
- 20 Tathe_UIUC_task5_1 50.6%
- 21 Hu_IOA_task5_2 49.9%
- 22 Tathe_UIUC_task5_2 49.8%
- 23 Hu_IOA_task5_1 49.6%
- 24 Huang_JAIST_task5_1 49.6%
- 25 Kim_SGU_task5_4 49.1%
- 26 Kim_SGU_task5_3 48.9%
- 27 Tathe_UIUC_task5_3 48.7%
- 28 Tathe_UIUC_task5_4 48.3%
- 29 Kim_SGU_task5_2 48%
- 30 Wu_XMU_task5_1 47.2%
- 31 Kim_SGU_task5_1 46.8%
- 32 ZC_Inst_task5_1 46%
- 33 Guan_HEU_task5_1 43.4%
- 34 Song_BIT_task5_1 43.3%
- 35 Song_BIT_task5_2 41.5%
- 36 Song_BIT_task5_3 41%
- 37 Xu_HUST_task5_1 31.9%
Source: Official ADQA-Bench leaderboard
Source: Official ADQA-Bench leaderboard
| # | Entry | Score | Setting | Submitted by | Date | Link | Source |
|---|---|---|---|---|---|---|---|
| 1 | Lim_CAU_task5_4 96B | 58.3% | Chung-Ang University | dcase.community | Official leaderboard | ||
| 2 | Lim_CAU_task5_3 80B | 58.1% | Chung-Ang University | dcase.community | Official leaderboard | ||
| 3 | Lim_CAU_task5_1 8B | 57.3% | Chung-Ang University | dcase.community | Official leaderboard | ||
| 4 | Nam_IND_task5_2 60B | 57.2% | Independent researcher | dcase.community | Official leaderboard | ||
| 5 | Nam_IND_task5_4 94B | 57.1% | Independent researcher | dcase.community | Official leaderboard | ||
| 6 | Lim_CAU_task5_2 8B | 57.1% | Chung-Ang University | dcase.community | Official leaderboard | ||
| 7 | Hu_IOA_task5_4 8.6B | 57% | Institute of Acoustics, Chinese Academy of Sciences | dcase.community | Official leaderboard | ||
| 8 | Nam_IND_task5_3 60B | 56.7% | Independent researcher | dcase.community | Official leaderboard | ||
| 9 | Hu_IOA_task5_3 8.6B | 56.7% | Institute of Acoustics, Chinese Academy of Sciences | dcase.community | Official leaderboard | ||
| 10 | Yin_XJTLU_task5_1 30B | 56% | Xi'an Jiaotong-Liverpool University | dcase.community | Official leaderboard | ||
| 11 | Nam_IND_task5_1 60B | 55.9% | Independent researcher | dcase.community | Official leaderboard | ||
| 12 | Yin_XJTLU_task5_4 8B | 55.8% | Xi'an Jiaotong-Liverpool University | dcase.community | Official leaderboard | ||
| 13 | Yin_XJTLU_task5_2 8B | 55.6% | Xi'an Jiaotong-Liverpool University | dcase.community | Official leaderboard | ||
| 14 | Yin_XJTLU_task5_3 8B | 55.3% | Xi'an Jiaotong-Liverpool University | dcase.community | Official leaderboard | ||
| 15 | Cheng_Surrey_task5_1 8B | 53.9% | University of Surrey | dcase.community | Official leaderboard | ||
| 16 | Cheng_Surrey_task5_2 8B | 53.5% | University of Surrey | dcase.community | Official leaderboard | ||
| 17 | | 53.2% | DCASE 2026 Task 5 Organizers | Sep 2025 | Official leaderboard | ||
| 18 | Zhang_WHU_task5_1 8B | 51.6% | Wuhan University + The Chinese University of Hongkong, Shenzhen | dcase.community | Official leaderboard | ||
| 19 | Zhang_WHU_task5_2 8B | 51.1% | Wuhan University + The Chinese University of Hongkong, Shenzhen | dcase.community | Official leaderboard | ||
| 20 | Tathe_UIUC_task5_1 8.9B | 50.6% | University of Illinois Urbana-Champaign + Carnegie Mellon University, WavLab | dcase.community | Official leaderboard | ||
| 21 | Hu_IOA_task5_2 7.6B | 49.9% | Institute of Acoustics, Chinese Academy of Sciences | dcase.community | Official leaderboard | ||
| 22 | Tathe_UIUC_task5_2 8.9B | 49.8% | University of Illinois Urbana-Champaign + Carnegie Mellon University, WavLab | dcase.community | Official leaderboard | ||
| 23 | Hu_IOA_task5_1 7.6B | 49.6% | Institute of Acoustics, Chinese Academy of Sciences | dcase.community | Official leaderboard | ||
| 24 | Huang_JAIST_task5_1 8B | 49.6% | Japan Advanced Institute of Science and Technology | dcase.community | Official leaderboard | ||
| 25 | Kim_SGU_task5_4 7B | 49.1% | Sogang University | dcase.community | Official leaderboard | ||
| 26 | Kim_SGU_task5_3 7B | 48.9% | Sogang University | dcase.community | Official leaderboard | ||
| 27 | Tathe_UIUC_task5_3 8.9B | 48.7% | University of Illinois Urbana-Champaign + Carnegie Mellon University, WavLab | dcase.community | Official leaderboard | ||
| 28 | Tathe_UIUC_task5_4 8.9B | 48.3% | University of Illinois Urbana-Champaign + Carnegie Mellon University, WavLab | dcase.community | Official leaderboard | ||
| 29 | Kim_SGU_task5_2 7B | 48% | Sogang University | dcase.community | Official leaderboard | ||
| 30 | Wu_XMU_task5_1 7B | 47.2% | Xiamen University + Tsinghua University | dcase.community | Official leaderboard | ||
| 31 | Kim_SGU_task5_1 7B | 46.8% | Sogang University | dcase.community | Official leaderboard | ||
| 32 | ZC_Inst_task5_1 8B | 46% | hangzhou dianzi university | dcase.community | Official leaderboard | ||
| 33 | Guan_HEU_task5_1 8B | 43.4% | Harbin Engineering University + University of Technology Sydney | dcase.community | Official leaderboard | ||
| 34 | Song_BIT_task5_1 7B | 43.3% | Beijing Institute of Technology | dcase.community | Official leaderboard | ||
| 35 | Song_BIT_task5_2 7B | 41.5% | Beijing Institute of Technology | dcase.community | Official leaderboard | ||
| 36 | Song_BIT_task5_3 7B | 41% | Beijing Institute of Technology | dcase.community | Official leaderboard | ||
| 37 | Xu_HUST_task5_1 5.4B | 31.9% | Huazhong University of Science and Technology | dcase.community | Official leaderboard |
| # | Model | Score | Setting | Submitted by | Date | Link | Source |
|---|---|---|---|---|---|---|---|
| 1 | | 53.2% | DCASE 2026 Task 5 Organizers | Sep 2025 | Official leaderboard |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official ADQA-Bench 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
- 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 14 months ago, in September 2025, for Qwen3-Omni-30B-A3B. 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 (37 entries).
- Pass Models scored 1 model we track, plus 36 other entries on its leaderboard.
- Warning Newest score The newest came 14 months ago, in September 2025, for Qwen3-Omni-30B-A3B.
- Note Frontier models None of its models are on the capabilities index.
- Pass Headroom The best is 53.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on dcase.community: 1 row for models we track and 36 other entries 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.