The Claude Code team stated that the effort value mapping test did not affect model performance.
Claude Code team members stated that recent testing changed the mapping method between internal values and effort levels, but the effort levels selected by users remained unchanged, and internal evaluations did not find any impact on model performance.
Thariq, a member of the Claude Code team, said that the team sometimes tests the API service configuration before a full rollout.The current test has changed the mapping between some internal values and effort levels, so Claude may display different values to some users.
The team stated that these values are not performance scores from 0 to 100, and the effort level selected by the user will still be applied normally. Internal evaluations also did not find any impact on model performance. The Claude Code official documentation also states that the underlying values for effort levels are calibrated separately for each model.
Currently, the conclusion that "performance is unaffected" is still based on internal team evaluation. If users encounter reproducible quality degradation, the team recommends submitting cases through the feedback function.
refer to:Team Description；Claude Code Model Configuration Documentation