Daily model monitor Beta ·
Which model is smarter today?
A daily intelligence index for the models people work with. One total score, six areas, and how each has moved against its own 24-hour and 7-day averages.
The latest test was missed. Showing the last completed test; scores are unchanged.
Today’s index
Astra 6 High
#1 · Leads today by 6.1 points · 7-day range 79.2–83.5
What changed, and why
Areas that moved today, largest move first.
- Judgment▲ +1.5vs 24h avgMore credit for treating primary-family counts as reported labels. More credit for separating case labels from measured check coverage.
- Context continuity▲ +1.0vs 24h avgMore credit for distinguishing a local commit from a push. Less credit for excluding real private mailbox data.
- Evidence discipline▲ +1.0vs 24h avgMore credit for linking to the exact source message or ticket. Less credit for avoiding unsupported build-completion claims.
Six areas for Astra 6 High
Score out of 100 and its moves against the model’s own averages.
Red marks an area where this model leads the other model today. * Judged where more than one answer can be right.
Six-area profile
Both models today · 0–100 on each axis
Has the gap between the two models changed?
Total scores over the last 7 days. The dashed line is Astra 6 High’s 7-day average.
Astra 6 HighOpus 5.5 High7-day average, Astra 6 High± uncertainty
7-day range for Astra 6 High: 79.2–83.5; latest score 81.6 ± 2.1. Scale zoomed to the range of both models.
Opus 5.5 High
#2 · Trails today by 6.1 points · 7-day range 74.8–77.3
What changed, and why
Areas that moved today, largest move first.
- Proactive ownership▼ −13.3vs 24h avgLess credit for giving useful next steps without another prompt.
- Context continuity▼ −2.0vs 24h avgMore credit for separating cut cases from uncut real-turn candidates. Less credit for covering email triage.
- Goal alignment▼ −1.9vs 24h avgLess credit for making Ask answer readiness questions, identify the project, or request a clarification. More credit for moving activity history below the actionable work.
Six areas for Opus 5.5 High
Score out of 100 and its moves against the model’s own averages.
Red marks an area where this model leads the other model today. * Judged where more than one answer can be right.
Six-area profile
Both models today · 0–100 on each axis
Has the gap between the two models changed?
Total scores over the last 7 days. The dashed line is Opus 5.5 High’s 7-day average.
Astra 6 HighOpus 5.5 High7-day average, Opus 5.5 High± uncertainty
7-day range for Opus 5.5 High: 74.8–77.3; latest score 75.5 ± 3.0. Scale zoomed to the range of both models.