GPT-Audio 1.5: benchmark results
Provider: OpenAI. Released 2026-02-23. Access: API.
Aggregate Rating 1673 ± 38, rank #490 of 10916 rated models, from 33 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MSI-Bench - English Background Speech Retrieval | 14.6 | All-pass rate (%; the 96 background-speech-retrieval cases, | 89.3 |
| MSI-Bench - Mandarin Sequential Constraints | 51 | All-pass rate (%; the 96 sequential-constraint cases, where | 82.1 |
| MusTBENCH - Transition-Aware Description | 23.6 | Transition-Aware Description (TAD, 208 questions): the model | 81.8 |
| MSI-Bench - English Sequential Constraints | 47.9 | All-pass rate (%; the 96 sequential-constraint cases, where | 78.6 |
| MSI-Bench - English Speaker Authority | 78.1 | All-pass rate (%; the 96 speaker-authority cases, where an u | 78.6 |
| MSI-Bench - Mandarin Scope Tracking | 45.8 | All-pass rate (%; the 96 scope-tracking cases, where group-w | 78.6 |
| VocalCoachBench - Triplet Ranking | 71.7 | Pairwise accuracy (%; for three performances of the same son | 77.8 |
| MSI-Bench - Mandarin Selective Disclosure | 39.6 | All-pass rate (%; the 96 selective-disclosure cases, where e | 75 |
| MSI-Bench - English | 49.5 | All-pass rate (%; all 576 English test cases over six multi- | 71.4 |
| MSI-Bench - English Constraint Prioritization | 55.2 | All-pass rate (%; the 96 constraint-prioritization cases, wh | 71.4 |
| MSI-Bench - English Scope Tracking | 37.5 | All-pass rate (%; the 96 scope-tracking cases, where group-w | 71.4 |
| MSI-Bench - English Selective Disclosure | 63.5 | All-pass rate (%; the 96 selective-disclosure cases, where e | 71.4 |
Interactive version: theaggregate.ai/model?slug=gpt-audio-1-5 · How It Works · Data refreshed daily, snapshot 2026-11-08.