Claude Opus 4.7 costs 20–30% more per session | Hacker News

Hacker News users debated Claude Opus 4.7’s reported 20–30% higher per-session cost.

Many questioned whether the model’s real-world quality gains justify the extra token spend.

The thread also reflected concern that Anthropic may be prioritizing profit over its stated mission.

More broadly, the discussion underscored a growing AI pricing question: when do higher model costs stop delivering clear ROI?