GPT-4o Audio (Preview 2025-06-03): benchmark results
Provider: OpenAI. Access: API.
Aggregate Rating 1556 ± 41, rank #1617 of 10916 rated models, from 12 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SEAL Showdown | 1271.9 | Arena Score | 100 |
| SonicBench - Counting | 67 | Accuracy (%; number of sound events (PicoAudio); 100 recogni | 83.9 |
| SonicBench - Brightness | 73 | Accuracy (%; spectral brightness (NSynth); 100 recognition a | 80.6 |
| SonicBench - Pitch | 71 | Accuracy (%; pitch height (NSynth notes, one-semitone contra | 80.6 |
| SonicBench - Texture | 90 | Accuracy (%; sound texture (ESC-50); 100 recognition and 100 | 75 |
| SonicBench | 59 | Accuracy (%; mean over 12 physical attributes of 100 recogni | 72.6 |
| SonicBench - Loudness | 52 | Accuracy (%; loudness (integrated LUFS contrasts of at least | 50 |
| AudioMC | 25.44 | Score (imported) | 48.1 |
| SEAL - AudioMultiChallenge - Text Output | 25.44 | Score | 45 |
| SonicBench - Reverberation | 50 | Accuracy (%; reverberation (SRIR room impulse responses); 10 | 25.8 |
| SonicBench - Duration | 49 | Accuracy (%; sound-event duration (TAU); 100 recognition and | 21 |
| SonicBench - Timbre | 46 | Accuracy (%; instrument timbre (NSynth); 100 recognition and | 3.2 |
Interactive version: theaggregate.ai/model?slug=gpt-4o-audio-preview-2025-06-03 · How It Works · Data refreshed daily, snapshot 2026-11-08.