Anthropic splits Claude subscribers into two tiers on Fable 5 as the five-week promotional limbo ends July 20. Max and Team Premium plans lock in permanent access to the company's flagship model at 50% of standard weekly usage limits, while Pro and Team Standard subscribers lose bundled access entirely and receive a one-time $100 usage credit instead. The split, confirmed in Anthropic's July 17 announcement via its @claudeai account on X, resolves a stretch of deadline chaos that saw the company push back the cutoff date three times.
Anthropic originally scheduled Fable 5's subscription exit for July 7, extended to July 12 after user backlash, then pushed again to July 19 as competitive dynamics shifted. For Pro and Team Standard users, the $100 credit buys less than it sounds like. At Fable 5's API rate of $50 per million output tokens, a single session producing 2 million output tokens (modest for an autonomous code migration or document analysis run) consumes the entire credit.
Anthropic has not said when the credit will be distributed or what a typical user is expected to get from it. After the credit is exhausted, continued access requires paying $10 per million input tokens and $50 per million output tokens, the steepest pricing for any Claude model in general availability.
Max and Team Premium subscribers get permanence but with a real reduction. The 50% cap on weekly usage arrives the same day the 50% boost to Claude Code weekly rate limits expires, effectively shrinking usable capacity from the promotional period.
Fable 5's path to this outcome traces back to the model's architecture. Part of Anthropic's Mythos class introduced above Opus in June 2026, Fable 5 uses adaptive reasoning that demands substantially more compute per token than Opus 4.8. The model carries a 1-million-token context window and 128,000-token maximum output, specs designed for multi-hour autonomous sessions that consume tokens at scale.
A mandatory 30-day data retention policy on all Mythos-class traffic adds operational overhead Opus models don't carry. The competitive pressure that shaped Anthropic's decision came from OpenAI. GPT-5.6 Sol reached general availability July 9 at $5 per million input tokens and $30 per million output tokens, half Fable 5's input cost and 60% of its output cost.
On the independent Artificial Analysis Intelligence Index, Fable 5 scores approximately 59.9 and Sol scores 58.9, a statistical tie on general capability. Sol completes equivalent tasks at roughly one-third Fable 5's per-task cost because it emits fewer output tokens per result, per Artificial Analysis.
Anthropic's pricing ladder after July 20 spans an order of magnitude: Haiku 4.5 at $1/$5 per million tokens, Sonnet 5 at $2/$10 (introductory pricing through August 31), Opus 4.8 at $5/$25, and Fable 5 at $10/$50. The 2:1 output cost gap between Fable 5 and Opus 4.8 is the largest within Anthropic's lineup.
Whether Fable 5 eventually returns as a standard Pro-tier benefit depends on how fast Anthropic can scale infrastructure to meet demand. A Claude Code lead engineer promised publicly in early July that the company aims to restore Fable 5 to subscriptions as soon as capacity allows.
July 20 is not that date.
