OpenAI
GPT-5.6 Sol benchmark results
11 results on 7 benchmarks from 4 sources.
Also listed as: gpt-5.6-sol-xhigh · gpt-5.6-sol_none · GPT-5.6 Sol (none) · gpt-5.6-sol_low · GPT-5.6 Sol (low) · gpt-5.6-sol_max · GPT-5.6 Sol (max) · GPT-5.6-sol · GPT-5.6 Sol (XHigh) · GPT-5.6 Sol (Max)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1485 | 20 | Overall, style control on; 38490 votes; 95% CI 1481–1489“gpt-5.6-sol-xhigh” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1285 | 18 | Overall, style control on; 10421 votes; 95% CI 1277–1292“gpt-5.6-sol-xhigh” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 93.5% | 14 | Epoch AI run, accuracy; effort: max | Independent run | Jul 9, 2026 |
| GPQA Diamond | 89.9% | 42 | Epoch AI run, accuracy; effort: low | Independent run | Aug 7, 2026 |
| GPQA Diamond | 82.8% | 101 | Epoch AI run, accuracy; effort: none | Independent run | Aug 7, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 100.0% | 1 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Jul 9, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 95.6% | 43 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: low | Independent run | Aug 7, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 68.9% | 131 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: none | Independent run | Aug 7, 2026 |
| ARC-AGI-2 | 92.5% | 5 | Semi-private set; $1.44/task; CoT“GPT-5.6 Sol (Max)” | Official leaderboard | Oct 22, 2026 |
| OSWorld 2.0 | 27.3% | 4 | Binary accuracy; 500 steps; reasoning: max; effort: max | Official leaderboard via Epoch AI | Jul 9, 2026 |
| τ²-bench | 46.9% | 4 | banking_knowledge domain, pass^1; effort: xhigh; submitted by Sierra; v1.0.1“GPT-5.6-sol” | Official leaderboard | Jul 22, 2026 |
FAQ
Where do these GPT-5.6 Sol scores come from?
From Epoch AI Benchmarking Hub, ARC Prize leaderboard, LMArena leaderboard dataset, and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.