Artificial Analysis says a top-five model has the highest win rate in six of seven AA-3D domains. In living bodies, sixth-ranked GPT-6.1 Sol scores 78.6%, tying Claude Sonnet 5.5 at 78.5%. The post says gaps under about 12 points are noise, and lists scene costs of $20.50 for Claude Opus 5.5, $18.11 for GPT-6 Astra, and $3.51 for GPT-6.1 Sol.
Part of a thread · earlier post
The same models top almost every AA-3D domain
A top-five model has the highest win rate in six of seven domains. In living bodies, sixth-ranked GPT-6.1 Sol (78.6%) ties with Claude Sonnet 5.5 (78.5%). Gaps under ~12 points are noise. Scenes by Claude Opus 5.5 ($20.50 a scene), GPT-6 Astra ($18.11), and GPT-6.1 Sol ($3.51) follow
Source: Artificial Analysis · x.comPublished · added here
