Glimmer Index - Starjar: leaderboard

Play Starjar under the source's paused-time physics driver, fixed v1 items and two-answer retry policy. The published score is relative to its reference policy and can exceed 100; it is not clipped or interpreted as a task pass rate. Completed six-game configurations retain exact route, effort, game commit, rules hash and driver metadata; partial configurations remain outside the comparison.

Metric: Score relative to reference policy (%; uncapped). Source: glimmerindex.com. Saturation forecast: Rough model projection: around 2026. 5 models tracked.

Top models

#ModelScoreOverall rank
1GPT-6 Astra (Glimmer Index Run: codex-gpt-6-astra-medium;model=openai/gpt-6-astra;provider=codex;route=gpt-6-astra;effort=medium;precision=unreported)112.3
2GPT-6.1 Terra (Glimmer Index Run: codex-gpt-6-1-terra-medium;model=openai/gpt-6.1-terra;provider=codex;route=gpt-6.1-terra;effort=medium;precision=unreported)103.7
3GPT-6.1 Sol (Glimmer Index Run: codex-gpt-6-1-sol-medium;model=openai/gpt-6.1-sol;provider=codex;route=gpt-6.1-sol;effort=medium;precision=unreported)101.6
4GPT-6 Luna (Glimmer Index Run: codex-gpt-6-luna-medium;model=openai/gpt-6-luna;provider=codex;route=gpt-6-luna;effort=medium;precision=unreported)95.2
5GPT-6 Sol (Glimmer Index Run: codex-gpt-6-sol-medium;model=openai/gpt-6-sol;provider=codex;route=gpt-6-sol;effort=medium;precision=unreported)90

Interactive version: theaggregate.ai/benchmark?slug=glimmer-index-starjar · How It Works · Data refreshed daily, snapshot 2026-11-05.