Leaderboard / Model
GPT-5.4 Pro
OpenAI · Closed weights · Released Mar 5, 2026 · $30 in · $180 out per 1M tokens
n/aOverall · not enough data
Score by job
Coding
Not enough data · 0 of 5 benchmarks
Agents
Not enough data · 0 of 4 benchmarks
Reasoning
77 · rank 8 · 4 of 4 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| GPQA Diamond | 94.6% | gpt-5.4-pro-2026-03-05_xhigh | 1.23 |
| ARC-AGI-2 | 83.3% | gpt-5.4-pro-2026-03-05_xhigh | 0.89 |
| SimpleBench | 74.1% | gpt-5.4-pro-2026-03-05_unknown | 0.87 |
| Humanity’s Last Exam | 44.3% | gpt-5.4-pro-2026-03-05_unknown | 0.75 |
Math
Not enough data · 2 of 5 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| FrontierMath Tiers 1–3 | 82.5% | gpt-5.4-pro-2026-03-05_xhigh | 1.00 |
| FrontierMath Tier 4 | 58.5% | gpt-5.4-pro-2026-03-05_xhigh | 0.73 |
Writing
Not enough data · 0 of 2 benchmarks
Design
Not enough data · 0 of 1 benchmarks