GPT-5.4 Nano (xHigh): benchmark results
GPT-5.4 Nano evaluated at the xhigh reasoning-effort setting. Provider: OpenAI. Released 2026-03-17. Access: API.
Unified ELO 1628 ± 1, rank #325 of 1761 rated models, from 59 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA IFBench | 75.92 | Accuracy (%) | 94.5 |
| AA Terminal-Bench Hard | 42.42 | Accuracy (%) | 89.2 |
| AA CritPt | 9.25 | Accuracy (%) | 84.6 |
| Artificial Analysis Intelligence Index | 30.92 | Intelligence Index | 82.1 |
| LiveBench Table Join | 50.59 | Score | 81.1 |
| AA Long Context Reasoning | 76.67 | Accuracy (%) | 80.8 |
| AA Humanity's Last Exam | 28.27 | Accuracy (%) | 80 |
| AA Omniscience - Software Engineering (SWE) | 38.1 | Accuracy (%) | 70.1 |
| AA GPQA Diamond | 81.72 | Accuracy (%) | 69.7 |
| AA Omniscience - Science, Engineering & Mathematics | 34.2 | Accuracy (%) | 68 |
| AA TAU-2 Bench | 76.02 | Accuracy (%) | 67.8 |
| AA-LCR | 72 | Accuracy (self-reported) | 67.1 |
Interactive version: theaggregate.ai/model?slug=gpt-5-4-nano-xhigh · How It Works · Data refreshed daily, snapshot 2026-09-05.