Anthropic Restricts Opus Model Access on Claude Pro Behind Extra Usage Paywall

Summary

Anthropic has restructured access to its Opus model family on Claude Pro, requiring users to enable and purchase optional extra usage beyond their standard subscription to unlock Opus models. This tiered access model means that while Sonnet and Haiku models remain available at the base Claude Pro tier, users seeking the most capable Opus models must opt into additional paid usage. The change affects Claude Code users and represents a strategic shift toward usage-based monetization of Anthropic's most computationally intensive models. To support users navigating the new structure, Anthropic has published detailed documentation on three methods for switching between available models: the /model command for quick interactive selection, the --model flag for session-specific choices, and environment variables for persistent default configuration.

• This change represents a shift toward usage-based pricing that allows Anthropic to manage inference costs while maintaining accessibility of lower-tier models

Editorial Opinion

Anthropic's decision to gate Opus access behind optional extra usage purchases is a calculated monetization strategy that acknowledges the substantial computational costs of serving the most capable models. While this may frustrate users expecting unlimited Opus access with Pro, it strikes a practical balance between profitability and accessibility by keeping Sonnet and Haiku available at the base tier. The comprehensive documentation for model switching should ease the transition, though long-term user satisfaction will depend on competitive pricing of the extra usage tier.