GPT-5.4 Mini: benchmark results
OpenAI's smaller, faster GPT-5.4 tier for cost-efficient reasoning workloads. Provider: OpenAI. Released 2026-03-17. Access: API.
Unified ELO 1658 ± 1, rank #113 of 1392 rated models, from 285 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AGC-Bench - hypobench | 0.48 | Dataset z-score | 100 |
| CalBench | 149 | Excess (self-reported) | 100 |
| GACL - Battleship | 85.94 | Normalized Score (0-100) | 100 |
| TempGlitch | 52.4 | Acc. (1 FPS) (self-reported) | 100 |
| AGC-Bench - pun_eval | 1.4 | Dataset z-score | 97.6 |
| AGC-Bench - humor_transfer | 0.97 | Dataset z-score | 96.3 |
| LLM-as-a-Reviewer | 99.7 | Low (NeurIPS 2022) (self-reported) | 95.5 |
| AGC-Bench - pollux_creativity | 0.92 | Dataset z-score | 95.1 |
| SLR-Bench - Hard | 28 | Accuracy (%) | 94.5 |
| Roboflow Vision Evals - Visual Understanding | 77.61 | Pass Rate (%) | 93.4 |
| Chartographer | 94 | ChartQA OA (self-reported) | 92.9 |
| AGC-Bench - pron_vs_prompt | 1.06 | Dataset z-score | 92.6 |
Interactive version: theaggregate.ai/model?slug=gpt-5-4-mini · How It Works · Data refreshed daily, snapshot 2026-09-05.