Anthropic pushes back on Sonnet 6 regression claims, says it's a flatly better model

pvncher · x · 2026-09-23

Responding to widespread claims that Sonnet 6 regressed vs 5.6, Anthropic team member pvncher says the new model should be "a flatly better model despite what the evals say," and urges users to report behavior issues via /feedback, noting the team is monitoring submissions closely and responding nimbly.

Original post →

More from Models

Models channel →