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

Re:Anthropic appears to be A/B testing reduced effort levels... by KronisLV on Saturday August 22, 2026 @03:26PM (#49402977)

> something started changing around March.

The economics catching up with the providers in regards to how much compute they can burn per request and have it make sense for them financially?

A sort of model collapse where Opus 5 seems to love throwing out long paragraphs of text and it needs to be "fixed" by changing the output style and other patches.

I'm not sure, it might also catch up to Kimi K3 and GLM 5.3 and the models that I'm moving to from Anthropic.

[ Parent | on HN ]

Re:Anthropic appears to be A/B testing reduced effort levels... by conception on Saturday August 22, 2026 @07:10PM (#49404755)

Inference is profitable though.

[ Parent | on HN ]

Re:Anthropic appears to be A/B testing reduced effort levels... by jazzyjackson on Saturday August 22, 2026 @07:38PM (#49404905)

Even that being the case, if the providers can squeeze more happy customers onto existing capacity they would likely act to increase profitability, no?

[ Parent | on HN ]

Re:Anthropic appears to be A/B testing reduced effort levels... by flyinglizard on Sunday August 23, 2026 @04:02AM (#49406929)

I pay $200 a month at home for roughly the same amount of tokens I pay $3k for (or possibly more) at work. I doubt they are both profitable and sustainable.

[ Parent | on HN ]