Key Info

Requesty measured the share of tokens routed to six new models versus their predecessors during the first 72 hours on its gateway, with adoption rates ranging from 5% to 49%.

Highlights

  • GPT-6 Sol saw the fastest adoption at 49% of tokens within 72 hours
  • Claude Fable 5.1 reached 43%, while GPT-6 Luna hit 35%
  • Claude Opus 5.5 and DeepSeek V4.1 Flash both reached 23%
  • Claude Sonnet 5.5 lagged significantly at just 5% adoption