GPT-6 Astra Pro: benchmark results
OpenAI GPT-6 Astra in Pro reasoning mode. Kept separate from standard Astra in benchmark results; launch announced September 3, 2026 with staged availability. Provider: OpenAI. Released 2026-09-03. Access: API.
Aggregate Rating 1965 ± 7, rank #4 of 10556 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MineBench | 2206 | Elo Rating | 100 |
| WeirdML | 93.57 | Average Score | 100 |
| Arabic Broad Leaderboard - Translation (incl Dialects) | 8.17 | Average Score (0-10) | 98.8 |
| Chess Puzzles (Epoch AI) | 76 | Accuracy (%) | 98.8 |
| OpenRouter GPQA Diamond | 95.5 | Accuracy (%) | 98.7 |
| Arabic Broad Leaderboard - MMLU | 9.92 | Average Score (0-10) | 98.4 |
| Arabic Broad Leaderboard - Trust & Safety | 10 | Average Score (0-10) | 98 |
| Last Translation Benchmark | 45.33 | Verified Translations (%) | 98 |
| Arabic Broad Leaderboard | 9.2 | Average Score (0-10) | 96.1 |
| SimpleBench | 86.5 | Score (AVG@5) | 95.5 |
| The Aggregate - Trip Outcome Prediction (5 Questions) | 69.8 | Accuracy (%) | 95.2 |
| Arabic Broad Leaderboard - Writing (incl Dialects) | 7.96 | Average Score (0-10) | 93.7 |
Interactive version: theaggregate.ai/model?slug=gpt-6-astra-pro · How It Works · Data refreshed daily, snapshot 2026-11-17.