VocalGrad
Models decide which way an acoustic property changes within a recording. The authors compare model accuracy with human listeners on the same binary questions.
- Made by
- The University of Tokyo and Tohoku University
- Released
- October 2026
- Kind
- Audio understanding and voice
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech
- Strongest model tested
-
Gemini 3 Flash #60 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
Kimi-Audio-7B-Instruct leads with 51.8% among models released by December 2025 Kimi-Audio-7B-Instruct leads with 51.8% among models released by December 2025
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. Every run is in the next chart. 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 December 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 Kimi-Audio-7B-Instruct Default prompt 51.8%
- 2 Step-Audio 2 mini Four-shot 51%
- 3 MiMo-Audio-7B-Instruct Default prompt 50.3%
- 4
Gemini 3 Flash Default prompt 50.2%
- 5
Audio Flamingo 3 swap_order 50%
Source: The VocalGrad paper
- 1 Kimi-Audio-7B-Instruct Default prompt 51.8%
- 2 MiMo-Audio-7B-Instruct Default prompt 50.3%
- 3
Gemini 3 Flash Default prompt 50.2%
- 4
Audio Flamingo 3 swap_order 50%
Source: The VocalGrad paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi-Audio-7B-Instruct | 51.8% | Default prompt | Apr 2025 | Official leaderboard |
| 2 | Step-Audio 2 mini | 51% | Four-shot | Official leaderboard | |
| 3 | MiMo-Audio-7B-Instruct | 50.3% | Default prompt | Sep 2025 | Official leaderboard |
| 4 | | 50.2% | Default prompt | Dec 2025 | Official leaderboard |
| 5 | | 50% | swap_order | Jul 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi-Audio-7B-Instruct | 51.8% | Default prompt | Apr 2025 | Official leaderboard |
| 2 | MiMo-Audio-7B-Instruct | 50.3% | Default prompt | Sep 2025 | Official leaderboard |
| 3 | | 50.2% | Default prompt | Dec 2025 | Official leaderboard |
| 4 | | 50% | swap_order | Jul 2025 | Official leaderboard |
Rankings by part
We read 13 parts from the benchmark’s paper in 2 groups: category and trajectory. Each ranks models on one part of the test. Kimi-Audio-7B-Instruct leads 9 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
speed
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 53.1%
- 2 MiMo-Audio-7B-Instruct 50.7%
- 3
Audio Flamingo 3 ↑1 50%
- 4
Gemini 3 Flash ↓1 50%
pitch
4 models we track have a score here.
- 1
Gemini 3 Flash ↑2 51.5%
- 2 Kimi-Audio-7B-Instruct ↓1 50.5%
- 3 MiMo-Audio-7B-Instruct ↓1 50.5%
- 4
Audio Flamingo 3 50%
volume
4 models we track have a score here.
- 1 MiMo-Audio-7B-Instruct ↑1 52.9%
- 2 Kimi-Audio-7B-Instruct ↓1 51.1%
- 3
Audio Flamingo 3 ↑1 50%
- 4
Gemini 3 Flash ↓1 49.9%
distortion
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 51.6%
- 2 MiMo-Audio-7B-Instruct 51.4%
- 3
Audio Flamingo 3 ↑1 50%
- 4
Gemini 3 Flash ↓1 48.1%
roughness
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 62%
- 2 MiMo-Audio-7B-Instruct 52%
- 3
Audio Flamingo 3 ↑1 50%
- 4
Gemini 3 Flash ↓1 44.8%
noise
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 56.5%
- 2
Gemini 3 Flash ↑1 55.5%
- 3 MiMo-Audio-7B-Instruct ↓1 53.2%
- 4
Audio Flamingo 3 50%
echo
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 54.5%
- 2
Gemini 3 Flash ↑1 50.3%
- 3 MiMo-Audio-7B-Instruct ↓1 50.2%
- 4
Audio Flamingo 3 50%
clarity
4 models we track have a score here.
- 1
Gemini 3 Flash ↑2 54%
- 2 Kimi-Audio-7B-Instruct ↓1 53.6%
- 3 MiMo-Audio-7B-Instruct ↓1 52.2%
- 4
Audio Flamingo 3 50%
vibration
4 models we track have a score here.
- 1 MiMo-Audio-7B-Instruct ↑1 50.3%
- 2 Kimi-Audio-7B-Instruct ↓1 50.1%
- 3
Audio Flamingo 3 ↑1 50%
- 4
Gemini 3 Flash ↓1 47.6%
jump
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 52.8%
- 2 MiMo-Audio-7B-Instruct 51.6%
- 3
Gemini 3 Flash 50.5%
- 4
Audio Flamingo 3 49.7%
linear
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 51.3%
- 2 MiMo-Audio-7B-Instruct 49.9%
- 3
Gemini 3 Flash 49.9%
- 4
Audio Flamingo 3 49.6%
quad first flat
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 50.1%
- 2
Audio Flamingo 3 ↑2 49.5%
- 3
Gemini 3 Flash 49.4%
- 4 MiMo-Audio-7B-Instruct ↓2 49.3%
quad last flat
4 models we track have a score here.
- 1 Kimi-Audio-7B-Instruct 53%
- 2
Gemini 3 Flash ↑1 50.9%
- 3 MiMo-Audio-7B-Instruct ↓1 50.4%
- 4
Audio Flamingo 3 49.7%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The VocalGrad paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Kimi-Audio-7B-Instruct | Default prompt | 51.8% | |||
| Kimi-Audio-7B-Instruct | swap_order | 51.4% | |||
| Kimi-Audio-7B-Instruct | audio_ref | 51.1% | |||
| MiMo-Audio-7B-Instruct | Default prompt | 50.3% | |||
| Kimi-Audio-7B-Instruct | Four-shot | 50.2% | |||
| Default prompt | 50.2% | ||||
| swap_order | 50% | ||||
| audio_ref | 50% | ||||
| MiMo-Audio-7B-Instruct | Four-shot | 49.9% | |||
| MiMo-Audio-7B-Instruct | audio_ref | 49.8% | |||
| Default prompt | 49.6% | ||||
| MiMo-Audio-7B-Instruct | swap_order | 49% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The VocalGrad paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Kimi-Audio-7B-Instruct | Default prompt | 51.8% | |||
| Kimi-Audio-7B-Instruct | swap_order | 51.4% | |||
| Kimi-Audio-7B-Instruct | audio_ref | 51.1% | |||
| MiMo-Audio-7B-Instruct | Default prompt | 50.3% | |||
| Kimi-Audio-7B-Instruct | Four-shot | 50.2% | |||
| Default prompt | 50.2% | ||||
| swap_order | 50% | ||||
| audio_ref | 50% | ||||
| MiMo-Audio-7B-Instruct | Four-shot | 49.9% | |||
| MiMo-Audio-7B-Instruct | audio_ref | 49.8% | |||
| Default prompt | 49.6% | ||||
| MiMo-Audio-7B-Instruct | swap_order | 49% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The VocalGrad paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official VocalGrad 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 11 months ago, in December 2025, for Gemini 3 Flash. 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 benchmark’s paper (5 entries).
- Pass Models scored 4 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for Gemini 3 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3 Flash (#60).
- Pass Headroom The best is 51.8%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 12 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.