"Anthropic recognizes that users are reporting 'worsened responses' over the past month, tracing these reports to three separate changes affecting Claude Code, the Claude Agent SDK, and Claude Cowork. All three issues have now been resolved as of April 20 (version 2.1.116)."
"The Claude team has promised to move ahead 'differently' and has explained how they will ensure similar issues are much less likely to happen again, emphasizing that they 'never intentionally degrade our models.'"
"On March 4, we changed Claude Code's default reasoning effort from high to medium to reduce the very long latency some users were experiencing. This was the wrong tradeoff, and we reverted this change on April 7 after users told us they'd prefer to default to higher intelligence."
Anthropic is currently dealing with restricted access to the Mythos model and user complaints regarding Claude's performance. Developers have expressed a desire to revert to the previous Opus version due to issues with instruction interpretation. The company acknowledges reports of degraded responses linked to changes in Claude Code, the Claude Agent SDK, and Claude Cowork, which have since been resolved. Anthropic aims to improve its processes to prevent similar issues in the future, emphasizing that it does not intentionally degrade model performance.
Read at DevOps.com
Unable to calculate read time
Collection
[
|
...
]