GPT-6 Astra (Medium): benchmark results

Provider: OpenAI. Released 2026-09-03. Access: API.

Aggregate Rating 1927 ± 8, rank #28 of 2257 rated models, from 261 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AI Coding Daily (Codex CLI) - Python Code Quality18.5Python Code Quality (max 20) points, LLM-judged rubric score100
Computer Anthology Terminal Tasks (Terminus-2)67pass@1 (%)100
DistDebug-Bench - Block67.7Coverage (%)100
DistDebug-Bench - Line53.9Coverage (%)100
DistDebug-Bench - Mention82.8Coverage (%)100
DrivingBench100Course progress (%, best of up to 3 attempts in one chat)100
DrivingBench (First Attempt)48.82Course progress (%, first attempt)100
Maze-Bench94.99Mean Score100
OpenVLHarness Study - Base - Average60.27Published Task Score (%)100
OpenVLHarness Study - Base - BC-VL46.12Published Task Score (%)100
OpenVLHarness Study - Base - CountQA82Published Task Score (%)100
OpenVLHarness Study - Base - ERQA76.5Published Task Score (%)100

Interactive version: theaggregate.ai/model?slug=gpt-6-astra-medium · How It Works · Data refreshed daily, snapshot 2026-11-17.