GPT-5.6 Sol (xHigh) — benchmark results
GPT-5.6 Sol evaluated at the xhigh reasoning-effort setting. Provider: OpenAI. Released 2026-07-09. Access: API.
Unified ELO 2096 ± 39, rank #10 of 1776 rated models, from 63 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA Omniscience - Software Engineering (SWE) - Kotlin | 94 | Accuracy (%) | 100 |
| CADGenBench | 0.53 | Aggregate CAD Score | 100 |
| ErdosBench | 3.12 | Judge Average (0-4) | 100 |
| Opus Magnum Bench | 70.86 | Human-normalized score (%) | 100 |
| PM-LLM-Benchmark | 40.2 | Score | 100 |
| AA Omniscience - Software Engineering (SWE) - Go | 86 | Accuracy (%) | 99.8 |
| AA Omniscience - Humanities & Social Sciences | 57.7 | Accuracy (%) | 99.6 |
| AA-Omniscience Accuracy | 57.97 | Accuracy (%) | 99.6 |
| Artificial Analysis Intelligence Index | 57.65 | Intelligence Index | 99.6 |
| AA Omniscience - Software Engineering (SWE) - TypeScript | 91.11 | Accuracy (%) | 99.4 |
| ARC-AGI-1 | 97.5 | Accuracy (%) | 99.4 |
| ARC-AGI-2 | 90 | Accuracy (%) | 99.4 |
Interactive version: theaggregate.ai/model?slug=gpt-5-6-sol-xhigh · How the rankings work · Data refreshed daily, snapshot 2026-07-22.