Claude Fable 5 Decision Splits Max And Pro Users After A 27-Million-View Row

A pricing announcement does not usually draw 27 million views. Anthropic’s did. The post that finally settled six weeks of rolling deadlines around Claude Fable 5, published to the company’s Claude account on Saturday 18 July, has been seen more than 26.9 million times and gathered upward of 53,000 likes, an unusually loud reception for a paragraph about usage limits. The reason is that the decision it confirmed lands very differently depending on which plan you pay for.

The terms are settled. From 20 July, Fable 5 is included in all Max and Team Premium plans at 50 per cent of their usage limits, made standard rather than promotional. Pro and Team Standard subscribers keep the model only through metered usage credits, softened by a one-time 100 dollar credit. That reverses Anthropic’s original intention, which was to remove Fable from subscriptions altogether.

For Max subscribers the mood is relief. Developer Simon Willison headlined his write-up “Claude make Fable 5 permanent”, and several outlets framed the news as the end of the limbo, with the model now a fixed part of the top tiers rather than a promotion waiting to lapse. After three cliff-edge deadlines in six weeks, each pushed back five to seven days hours before it bit, a permanent answer is what many of these users had been asking for.

For Pro subscribers the reading is colder. The Decoder described Anthropic as slashing Fable limits and pushing Pro users toward API pricing, and the arithmetic is what stings. A single heavy session that produces two million output tokens exhausts the entire 100 dollar credit on its own, so for anyone using Fable seriously the credit lasts one to a few sessions rather than weeks. As one summary put it, for a Pro subscriber hoping the plan would eventually regain full Fable access, 20 July is Anthropic saying in writing that it will not. Reddit’s verdict during the earlier restriction wave was blunter: bringing Fable back and then halving the usage was “not a good look”.

The path to the decision was messy enough to leave scars. This was the fourth deadline the model had carried since its 9 June launch, and the pattern each time was the same: Anthropic set a date, users complained, and the date moved. A false alarm on 17 July, and a brief spell where the status page showed credits as required for everyone before the company corrected it, added to the sense of a rollout being improvised in public. Anthropic’s own follow-up acknowledged as much, telling users the episode “has been frustrating” and framing the 50 per cent standard as an attempt to give the heaviest-use plans more certainty, with a promise to keep investing in capacity.

The wider read is the one worth keeping. A billing note becoming one of the company’s most-watched posts shows how central access to a single frontier model has become to the way people now work. And the split it drew, comfort at the top and sums at the bottom, is really a statement about who Anthropic can afford to serve at full strength while capacity stays tight.