FrontierCode 1.1
- OpenAI
- Anthropic
Results
| Model | Score | Source | Human-verified |
|---|---|---|---|
| Claude Opus 5.522 Sep 2026 · Anthropic | 54.4% | Introducing Claude Opus 5.5max effort; the post's prose gives 54.6% at default (medium) effort | Not yet |
| GPT-6 Astra3 Sep 2026 · OpenAI | 53.3% | GPT-6 Astra: A new generation of intelligencethe cell carries footnote 8 | Not yet |
Benchmark
cognition.comHigher is better