Anthropic appears to be A/B testing reduced effort levels in Claude Code

A developer reports that Anthropic is running a server-side A/B test on Claude Code versions 2.1.236 and later, shrinking the effort scale so that 'high' now maps to 10 out of 100—the same value 'low' used to have. Older versions and Opus 5 are unaffected. The changelog doesn't mention this change, leaving users confused about degraded performance.