PARADE Audio
Models answer questions about occupations and social status using audio clips. The evaluation tests whether voice cues bias their answers.
The release month follows the dataset repository history.
- Made by
- UCSC VLAA and the AHELM team
- Released
- March 2025
- Kind
- Bias, fairness and politics
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Propensity
- Behavior
- Bias
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech
- Strongest model tested
-
Gemini 2.5 Flash May 2025 #114 on the capability index
- Last model added
- May 2025
- Scores refreshed
- November 20th, 2026
gpt-4o-mini-transcribe transcription (GPT-4o + gpt-4o-mini-transcribe) tops the leaderboard with 85.8% Qwen2.5-Omni-7B leads with 63.4% among models released by May 2025
Best score per entry from AHELM’s leaderboard, read on November 20th, 2026. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model on its own is Qwen2.5-Omni-7B, at 63.4%. Agents built around a model, such as gpt-4o-mini-transcribe transcription (GPT-4o + gpt-4o-mini-transcribe), are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.
The newest model with a score here came out in May 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
- 85.8 gpt-4o-mini-transcribe transcription (GPT-4o + gpt-4o-mini-transcribe)
- 85.8 gpt-4o-transcribe transcription (GPT-4o + gpt-4o-transcribe)
- 85.7 gpt-4o-mini-audio-preview-2024-12-17
- 85.7 whisper-1 transcription (GPT-4o + whisper-1)
- 84.7 gpt-4o-audio-preview-2024-10-01
- 77.9 gpt-4o-audio-preview-2024-12-17
- 63.4
Qwen2.5-Omni-7B
- 51.4
Gemini 2.5 Flash
- 46.5
Gemini 2.0 Flash
- 46.3
Gemini 2.0 Flash
- 43.6
Gemini 2.0 Flash-Lite
- 32.4
Gemini 2.5 Pro
- 31.2
Gemini 1.5 Flash
- 29.2
Gemini 1.5 Flash
- 21.7
Gemini 1.5 Pro
- 21.5
Gemini 1.5 Pro
- 20.9
Qwen2-Audio-7B-Instruct
- 1 gpt-4o-mini-transcribe transcription (GPT-4o + gpt-4o-mini-transcribe) Nov 2024 85.8%
- 2 gpt-4o-transcribe transcription (GPT-4o + gpt-4o-transcribe) Nov 2024 85.8%
- 3 gpt-4o-mini-audio-preview-2024-12-17 85.7%
- 4 whisper-1 transcription (GPT-4o + whisper-1) Nov 2024 85.7%
- 5 gpt-4o-audio-preview-2024-10-01 84.7%
- 6 gpt-4o-audio-preview-2024-12-17 77.9%
- 7
Qwen2.5-Omni-7B 63.4%
- 8
Gemini 2.5 Flash May 2025 51.4%
- 9
Gemini 2.0 Flash Dec 2024 46.5%
- 10
Gemini 2.0 Flash Feb 2025 46.3%
- 11
Gemini 2.0 Flash-Lite 43.6%
- 12
Gemini 2.5 Pro May 2025 32.4%
- 13
Gemini 1.5 Flash Sep 2024 31.2%
- 14
Gemini 1.5 Flash May 2024 29.2%
- 15
Gemini 1.5 Pro May 2024 21.7%
- 16
Gemini 1.5 Pro Sep 2024 21.5%
- 17
Qwen2-Audio-7B-Instruct 20.9%
Source: AHELM’s leaderboard
- 1
Qwen2.5-Omni-7B 63.4%
- 2
Gemini 2.5 Flash May 2025 51.4%
- 3
Gemini 2.0 Flash Dec 2024 46.5%
- 4
Gemini 2.0 Flash Feb 2025 46.3%
- 5
Gemini 2.0 Flash-Lite 43.6%
- 6
Gemini 2.5 Pro May 2025 32.4%
- 7
Gemini 1.5 Flash Sep 2024 31.2%
- 8
Gemini 1.5 Flash May 2024 29.2%
- 9
Gemini 1.5 Pro May 2024 21.7%
- 10
Gemini 1.5 Pro Sep 2024 21.5%
- 11
Qwen2-Audio-7B-Instruct 20.9%
Source: AHELM’s leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gpt-4o-mini-transcribe transcription (GPT-4o + gpt-4o-mini-transcribe) Nov 2024 | 85.8% | Nov 2024 | Official leaderboard | |
| 2 | gpt-4o-transcribe transcription (GPT-4o + gpt-4o-transcribe) Nov 2024 | 85.8% | Nov 2024 | Official leaderboard | |
| 3 | gpt-4o-mini-audio-preview-2024-12-17 | 85.7% | Official leaderboard | ||
| 4 | whisper-1 transcription (GPT-4o + whisper-1) Nov 2024 | 85.7% | Nov 2024 | Official leaderboard | |
| 5 | gpt-4o-audio-preview-2024-10-01 | 84.7% | Official leaderboard | ||
| 6 | gpt-4o-audio-preview-2024-12-17 | 77.9% | Official leaderboard | ||
| 7 | | 63.4% | Mar 2025 | Official leaderboard | |
| 8 | | 51.4% | May 2025 | Official leaderboard | |
| 9 | | 46.5% | Dec 2024 | Official leaderboard | |
| 10 | | 46.3% | Feb 2025 | Official leaderboard | |
| 11 | | 43.6% | Feb 2025 | Official leaderboard | |
| 12 | | 32.4% | May 2025 | Official leaderboard | |
| 13 | | 31.2% | Aug 2024 | Official leaderboard | |
| 14 | | 29.2% | May 2024 | Official leaderboard | |
| 15 | | 21.7% | May 2024 | Official leaderboard | |
| 16 | | 21.5% | Sep 2024 | Official leaderboard | |
| 17 | | 20.9% | Aug 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 63.4% | Mar 2025 | Official leaderboard | |
| 2 | | 51.4% | May 2025 | Official leaderboard | |
| 3 | | 46.5% | Dec 2024 | Official leaderboard | |
| 4 | | 46.3% | Feb 2025 | Official leaderboard | |
| 5 | | 43.6% | Feb 2025 | Official leaderboard | |
| 6 | | 32.4% | May 2025 | Official leaderboard | |
| 7 | | 31.2% | Aug 2024 | Official leaderboard | |
| 8 | | 29.2% | May 2024 | Official leaderboard | |
| 9 | | 21.7% | May 2024 | Official leaderboard | |
| 10 | | 21.5% | Sep 2024 | Official leaderboard | |
| 11 | | 20.9% | Aug 2024 | Official leaderboard |
Rankings by part
We read 4 parts from AHELM’s leaderboard, each ranking models on one part of the test, such as occupation, female voice, status, female voice and occupation, male voice. Qwen2.5-Omni-7B leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
occupation, female voice
11 models we track have a score here.
- 1
Qwen2.5-Omni-7B 44.4%
- 2
Gemini 2.5 Flash May 2025 34.4%
- 3
Gemini 2.0 Flash Dec 2024 28.7%
- 4
Gemini 2.0 Flash Feb 2025 28.7%
- 5
Gemini 2.0 Flash-Lite 28.5%
- 6
Gemini 2.5 Pro May 2025 15.2%
- 7
Qwen2-Audio-7B-Instruct ↑4 14.1%
- 8
Gemini 1.5 Pro May 2024 ↑1 8.94%
- 9
Gemini 1.5 Flash Sep 2024 ↓2 8.40%
- 10
Gemini 1.5 Pro Sep 2024 5.15%
status, female voice
11 models we track have a score here.
- 1
Qwen2.5-Omni-7B 83%
- 2
Gemini 2.0 Flash Dec 2024 ↑1 69%
- 3
Gemini 2.0 Flash Feb 2025 ↑1 68%
- 4
Gemini 2.5 Flash May 2025 ↓2 66%
- 5
Gemini 2.0 Flash-Lite 63%
- 6
Gemini 1.5 Flash May 2024 ↑2 55%
- 7
Gemini 2.5 Pro May 2025 ↓1 54%
- 8
Gemini 1.5 Flash Sep 2024 ↓1 53%
- 9
Gemini 1.5 Pro May 2024 35%
- 10
Gemini 1.5 Pro Sep 2024 35%
occupation, male voice
11 models we track have a score here.
- 1
Qwen2.5-Omni-7B 45%
- 2
Gemini 2.5 Flash May 2025 37.1%
- 3
Gemini 2.0 Flash Feb 2025 ↑1 27.6%
- 4
Gemini 2.0 Flash Dec 2024 ↓1 27.4%
- 5
Gemini 2.0 Flash-Lite 23%
- 6
Gemini 2.5 Pro May 2025 14.4%
- 7
Qwen2-Audio-7B-Instruct ↑4 11.4%
- 8
Gemini 1.5 Flash Sep 2024 ↓1 11.4%
- 9
Gemini 1.5 Pro May 2024 9.76%
- 10
Gemini 1.5 Pro Sep 2024 3.79%
status, male voice
11 models we track have a score here.
- 1
Qwen2.5-Omni-7B 81%
- 2
Gemini 2.5 Flash May 2025 68%
- 3
Gemini 2.0 Flash Dec 2024 61%
- 4
Gemini 2.0 Flash Feb 2025 61%
- 5
Gemini 2.0 Flash-Lite 60%
- 6
Gemini 1.5 Flash May 2024 ↑2 56%
- 7
Gemini 1.5 Flash Sep 2024 52%
- 8
Gemini 2.5 Pro May 2025 ↓2 46%
- 9
Gemini 1.5 Pro Sep 2024 ↑1 42%
- 10
Gemini 1.5 Pro May 2024 ↓1 33%
2 models have held number one on AHELM’s leaderboard since its release in March 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: AHELM’s leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: AHELM’s 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 Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6%
- Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts.
Gemini 4 Carbon 0%
- Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior.
Inkling 2 Mini 8.30%
- Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions.
Inkling 2 Small 4.70%
- Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior.
Inkling 2 Mini 9.80%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 18 months ago, in May 2025, for Gemini 2.5 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 official leaderboard (17 entries).
- Pass Models scored 11 models we track, plus 6 other entries on its leaderboard.
- Warning Newest score The newest came 18 months ago, in May 2025, for Gemini 2.5 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 2.5 Flash (#114).
- Pass Headroom The best is 63.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on AHELM: 14 rows for models we track and 3 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.