GPT-5.4 Mini (xHigh): benchmark results
GPT-5.4 Mini evaluated at the xhigh reasoning-effort setting. Provider: OpenAI. Released 2026-03-17. Access: API.
Aggregate Rating 1697 ± 10, rank #369 of 2168 rated models, from 168 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MLCR-AA - Conciseness | 99.44 | Conciseness (%) | 95.8 |
| AA Terminal-Bench Hard | 52.27 | Accuracy (%) | 95.7 |
| AA Global-MMLU-Lite - Burmese | 88.42 | Accuracy (%) | 93.6 |
| AA IFBench | 73.27 | Accuracy (%) | 90.6 |
| AA Global-MMLU-Lite - Swahili | 88.33 | Accuracy (%) | 90.5 |
| Korean CSAT 2026 (Easy Mode) - Physics I | 47 | Points (out of 50) | 88.5 |
| AA Global-MMLU-Lite - Chinese | 90.92 | Accuracy (%) | 87.3 |
| AA Global-MMLU-Lite - Bengali | 89.42 | Accuracy (%) | 86.4 |
| AA Global-MMLU-Lite - Indonesian | 90.83 | Accuracy (%) | 85.5 |
| AA Global-MMLU-Lite - Korean | 89.08 | Accuracy (%) | 85.5 |
| AA Global-MMLU-Lite | 89.41 | Accuracy (%) | 84.5 |
| AA Global-MMLU-Lite - Yoruba | 78 | Accuracy (%) | 84.5 |
Interactive version: theaggregate.ai/model?slug=gpt-5-4-mini-xhigh · How It Works · Data refreshed daily, snapshot 2026-10-20.