Updated
Updated · Anthropic · Jul 23
Anthropic Launches Claude Opus 5 at Half Fable 5 Cost
Updated
Updated · Anthropic · Jul 23

Anthropic Launches Claude Opus 5 at Half Fable 5 Cost

3 articles · Updated · Anthropic · Jul 23

Summary

  • Claude Opus 5 is available now as Anthropic’s new default on Claude Max, matching near-Fable 5 intelligence while keeping Opus 4.8 pricing at $5 per million input tokens and $25 per million output tokens.
  • Half-cost positioning is backed by benchmark gains: Opus 5 leads Frontier-Bench, scores within 0.5% of Fable 5 on CursorBench at max effort, and triples the next-best model on ARC-AGI 3.
  • Opus 5 also posted stronger enterprise and research results, including 10.2-point gains over Opus 4.8 on internal organic chemistry tasks and broad improvements in coding, automation, finance, legal and visual-output workflows.
  • Safety remains a constraint rather than a selling point: Anthropic said Opus 5 is its most aligned model yet, with a 2.3 misaligned-behavior score, but still trails Mythos 5 on offensive cybersecurity and advanced biology work.
  • Developers also get a Fast mode running about 2.5 times quicker, plus beta features for mid-conversation tool changes and automatic API fallbacks when safety filters flag requests.

Insights

Anthropic claims Opus 5 matches elite AI intelligence at half the cost, but what hidden workflow trade-offs are users actually paying for?
If Opus 5 deliberately trails Mythos 5 in cybersecurity, are specialized AI models becoming the new gatekeepers for high-risk digital threats?
With new evaluator loops and mid-task tool switching, is Opus 5 quietly transitioning Claude from a chatbot into an autonomous digital employee?