GPT-6 Luna (Medium): benchmark results
Provider: OpenAI. Released 2026-09-22. Access: API.
Aggregate Rating 1727 ± 9, rank #295 of 2178 rated models, from 203 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Mercury 3 Release - AA LCR v1.1 | 78.3 | Score (%) | 100 |
| Mercury 3 Release - AA Omniscience Public | 46.94 | Accuracy (%) | 100 |
| PLCC - History | 96 | Accuracy (%) | 94.2 |
| PLCC - Grammar | 93 | Accuracy (%) | 93.8 |
| PLCC - Overall | 91.83 | Mean category accuracy (%) | 91.2 |
| PLCC - Art & Entertainment | 88 | Accuracy (%) | 89.9 |
| PLCC - Geography | 96 | Accuracy (%) | 87.9 |
| PLCC - Culture & Tradition | 92 | Accuracy (%) | 86.4 |
| MLCR-AA - Accuracy | 79.44 | Accuracy (%) | 85.7 |
| PLCC - Vocabulary | 86 | Accuracy (%) | 85 |
| AA Long Context Reasoning | 78.33 | Accuracy (%) | 78.9 |
| Wolfram LLM Benchmarking Project | 57.1 | Correct Functionality (%) | 77.5 |
Interactive version: theaggregate.ai/model?slug=gpt-6-luna-medium · How It Works · Data refreshed daily, snapshot 2026-11-07.