Compare AIFind AIAI NewsAI Courses
About Us
PrivacyTermsFAQContactContact
AIB Inc.Company info
© 2026 AIB Inc.

Anthropic Tests Claude Code Effort Levels

Anthropic Tests Claude Code Effort Levels

twitter.com·Sunday, August 23, 2026
  • •Anthropic appears to test reduced effort levels in Claude Code 2.1.236+ sessions
  • •X post says the change is server-side, not part of the Claude Code app
  • •Older versions and Opus 5 are reportedly unaffected; Hacker News logged 155 score, 145 comments
  • •Anthropic appears to test reduced effort levels in Claude Code 2.1.236+ sessions
  • •X post says the change is server-side, not part of the Claude Code app
  • •Older versions and Opus 5 are reportedly unaffected; Hacker News logged 155 score, 145 comments
  • •Anthropic appears to test reduced effort levels in Claude Code 2.1.236+ sessions
  • •X post says the change is server-side, not part of the Claude Code app
  • •Older versions and Opus 5 are reportedly unaffected; Hacker News logged 155 score, 145 comments
  • •Anthropic appears to test reduced effort levels in Claude Code 2.1.236+ sessions
  • •X post says the change is server-side, not part of the Claude Code app
  • •Older versions and Opus 5 are reportedly unaffected; Hacker News logged 155 score, 145 comments

A Hacker News discussion on August 22 centered on a post from X user 🥔🥔🥔 saying Anthropic is running a server-side experiment in Claude Code, not changing the app itself. The post said Anthropic enrolls “fable 5” sessions on Claude Code 2.1.236+ into an experiment that shrinks the effort scale.

The same post said older Claude Code versions and Opus 5 are left alone. It described the change as probably an A/B test (comparing variants with different users), meaning not everyone will see it, and said users may notice “high” feeling like “low.” Hacker News listed Score: 155 and Comments: 145.

A Hacker News discussion on August 22 centered on a post from X user 🥔🥔🥔 saying Anthropic is running a server-side experiment in Claude Code, not changing the app itself. The post said Anthropic enrolls “fable 5” sessions on Claude Code 2.1.236+ into an experiment that shrinks the effort scale.

The same post said older Claude Code versions and Opus 5 are left alone. It described the change as probably an A/B test (comparing variants with different users), meaning not everyone will see it, and said users may notice “high” feeling like “low.” Hacker News listed Score: 155 and Comments: 145.

Read original (English)·Aug 22, 2026
Coding#anthropic#claude code#ab test#effort levels#opus 5#hacker news#server side#claude code 2 1 236