Updated
Updated · InfoWorld · Jul 27
Anthropic Launches Claude Opus 5 at $5 per Million Tokens, Matching Opus 4.8 Cost
Updated
Updated · InfoWorld · Jul 27

Anthropic Launches Claude Opus 5 at $5 per Million Tokens, Matching Opus 4.8 Cost

3 articles · Updated · InfoWorld · Jul 27

Summary

  • Claude Opus 5 became Anthropic’s new default model on Claude Max after its July 24 release, with the company positioning it as a stronger option for coding, knowledge work and AI agents.
  • At the same $5 per million input tokens and $25 per million output tokens as Opus 4.8, Anthropic said the model delivers sharply better performance rather than a price increase.
  • On Frontier-Bench v0.1, Anthropic said Opus 5 more than doubled Opus 4.8’s performance and surpassed rival models at a lower cost per task.
  • On CursorBench 3.2, Opus 5 scored within 0.5% of Claude Fable 5’s peak while costing half as much per task, underscoring Anthropic’s push to narrow frontier-model gaps more cheaply.
  • Alongside Opus 5, Anthropic also rolled out beta features for mid-conversation tool changes on the Claude Platform and automatic API fallbacks for developers.

Insights

Anthropic claims Opus 5 fixes deep bugs at half the cost, but will its token-hungry agentic loops secretly drain developer budgets?
With Opus 5 matching frontier models for science and coding at stagnant prices, are we witnessing the end of premium-priced AI dominance?
If Opus 5 automatically falls back to older models for safety, could this hidden downgrade silently compromise high-stakes enterprise workflows?