Benchmark Almanac

VocalGrad

Kimi-Audio-7B-Instruct 51.8% leads since October 2026
Not in model cards

Models decide which way an acoustic property changes within a recording. The authors compare model accuracy with human listeners on the same binary questions.

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
Audio, Text → Text
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
5
Entries in its paper, 4 models we track

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:

  1. 1 Kimi-Audio-7B-Instruct Default prompt 51.8%
  2. 2 Step-Audio 2 mini Four-shot 51%
  3. 3 MiMo-Audio-7B-Instruct Default prompt 50.3%
  4. 4 Gemini 3 Flash Default prompt 50.2%
  5. 5 Audio Flamingo 3 swap_order 50%

Source: The VocalGrad paper

#EntryScoreSettingDateSource
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 Gemini 3 Flash 50.2% Default prompt Dec 2025 Official leaderboard
5 Audio Flamingo 3 50% swap_order Jul 2025 Official leaderboard
#ModelScoreSettingDateSource
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 Gemini 3 Flash 50.2% Default prompt Dec 2025 Official leaderboard
4 Audio Flamingo 3 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. 1 Kimi-Audio-7B-Instruct 53.1%
  2. 2 MiMo-Audio-7B-Instruct 50.7%
  3. 3 Audio Flamingo 3 ↑1 50%
  4. 4 Gemini 3 Flash ↓1 50%

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.

Lowest effort to highest, light to dark
Kimi-Audio-7B-Instruct
Four-shot audio_ref swap_order Default pro…
51.8%
MiMo-Audio-7B-Instruct
swap_order audio_ref Default pro…
50.3%
Audio Flamingo 3
Default pro… swap_order
50%

Source: The VocalGrad paper

ModelEffortHarnessScoreCost per taskOutput 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%
Gemini 3 Flash Default prompt 50.2%
Audio Flamingo 3 swap_order 50%
Audio Flamingo 3 audio_ref 50%
MiMo-Audio-7B-Instruct Four-shot 49.9%
MiMo-Audio-7B-Instruct audio_ref 49.8%
Audio Flamingo 3 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.

Lowest effort to highest, light to dark
Kimi-Audio-7B-Instruct
Four-shot audio_ref swap_order Default pro…
51.8%
MiMo-Audio-7B-Instruct
swap_order audio_ref Default pro…
50.3%
Audio Flamingo 3
Default pro… swap_order
50%

Source: The VocalGrad paper

ModelEffortHarnessScoreCost per taskOutput 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%
Gemini 3 Flash Default prompt 50.2%
Audio Flamingo 3 swap_order 50%
Audio Flamingo 3 audio_ref 50%
MiMo-Audio-7B-Instruct Four-shot 49.9%
MiMo-Audio-7B-Instruct audio_ref 49.8%
Audio Flamingo 3 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.

Moonshot AI
Benchmark paper Not tested since Oct 2026

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.

Plot with other benchmarks
In the benchmark’s paper
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Kimi-Audio-7B-Instruct 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Kimi-Audio-7B-Instruct

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

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

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.

How we check