GPT-6 Astra Pro: benchmark results
Provider: OpenAI. Released 2026-09-03. Access: API.
Unified ELO 2011 ± 23, rank #7 of 1610 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MineBench | 2253 | Elo Rating | 100 |
| WeirdML | 93.57 | Average Score | 100 |
| OpenRouter GPQA Diamond | 95.5 | Accuracy (%) | 99.3 |
| Arabic Broad Leaderboard - MMLU | 9.92 | Average Score (0-10) | 98.6 |
| Arabic Broad Leaderboard - Translation (incl Dialects) | 8.17 | Average Score (0-10) | 98.6 |
| Arabic Broad Leaderboard - Trust & Safety | 10 | Average Score (0-10) | 98.2 |
| SimpleBench | 86.5 | Score (AVG@5) | 98.1 |
| Last Translation Benchmark | 45.33 | Verified Translations (%) | 97.7 |
| Arabic Broad Leaderboard | 9.2 | Average Score (0-10) | 95.5 |
| Arabic Broad Leaderboard - Writing (incl Dialects) | 7.96 | Average Score (0-10) | 94.5 |
| Arabic Broad Leaderboard - RAG QA | 8.46 | Average Score (0-10) | 86.4 |
| KernelBench Hub - Mega | 24.8 | Best Speedup vs Reference (x) | 85.7 |
Interactive version: theaggregate.ai/model?slug=gpt-6-astra-pro · How It Works · Data refreshed daily, snapshot 2026-10-03.