gpt-5.4-pro-2026-03-05
94.6%±1.6%
93.0%CI96.2%
aiqbench · LLM benchmarks
94.6%±1.6%
93.0%CI96.2%
94.0%±1.5%
92.5%CI95.5%
93.9%±1.6%
92.3%CI95.5%
One grid, every column. Pick axes in the Axes panel — one axis ranks like a single board, several blend into a weighted composite, all of them read as the index. Rank is assigned before any filter and never renumbers.
| Model | Org | ± CI | Swing | Add to compare | |||||
|---|---|---|---|---|---|---|---|---|---|
| 7 | gpt-5.4-pro-2026-03-05 | OpenAI | Mar 2026 | $30 / $180 | 94.6% | 93.0%–96.2% | #20–171 | ||
| 11 | gpt-5.5-pre-release | OpenAI | — | — | 94.0% | 92.5%–95.5% | #22–36 | ||
| 13 | gpt-5.5-pro-pre-release | OpenAI | — | — | 93.9% | 92.3%–95.5% | #23–60 | ||
| 15 | gpt-5.6-sol | OpenAI | Sep 2026 | $4 / $20 | 93.5% | 91.9%–95.1% | #6–25 | ||
| 17 | gpt-5.6-terra | OpenAI | Sep 2026 | $2 / $12 | 93.3% | 91.8%–94.8% | #13–59 | ||
| 18 | gpt-5.4-2026-03-05 | OpenAI | Sep 2026 | $2.50 / $15 | 93.3% | 91.5%–95.1% | #12–29 | ||
| 27 | gpt-5.6-luna | OpenAI | Sep 2026 | $0.20 / $1.20 | 91.6% | 89.9%–93.3% | #30–146 | ||
| 28 | gpt-5.2-2025-12-11 | OpenAI | Dec 2025 | $1.75 / $14 | 91.4% | 89.6%–93.2% | #34–92 | ||
| 36 | gpt-5.5 | OpenAI | Sep 2026 | $5 / $30 | 90.7% | 88.9%–92.4% | #9–34 | ||
| 58 | gpt-5.1-2025-11-13 | OpenAI | Nov 2025 | $1.25 / $10 | 87.6% | 85.8%–89.5% | #45–242 | ||
| 63 | gpt-5.4-mini-2026-03-17 | OpenAI | Oct 2026 | $0.75 / $4.50 | 86.9% | 84.5%–89.3% | #75–180 | ||
| 65 | gpt-5-2025-08-07 | OpenAI | Aug 2025 | $1.25 / $10 | 86.2% | 84.1%–88.3% | #64–149 | ||
| 81 | gpt-5.5-instant | OpenAI | Sep 2026 | — | 82.5% | 80.3%–84.8% | #52–114 | ||
| 91 | gpt-5.4-nano-2026-03-17 | OpenAI | Oct 2026 | $0.20 / $1.25 | 78.5% | 76.0%–80.9% | #95–222 | ||
| 101 | gpt-5-mini-2025-08-07 | OpenAI | Aug 2025 | $0.25 / $2 | 75.0% | 72.7%–77.3% | #126–343 | ||
| 109 | gpt-5-nano-2025-08-07 | OpenAI | Aug 2025 | $0.05 / $0.40 | 69.4% | 66.7%–72.2% | #132–399 |
Showing GPQA Diamond exactly as the source published it — its own score and confidence interval, no blending. Select a second axis to build a weighted composite.