Strongest AI Fails to Sell? Anthropic's Flagship Model Overtaken by Its Own Budget Version

The AI industry's long-held assumption that 'stronger models command higher enterprise spending' is now being challenged. Anthropic's most powerful and highest-priced model, Fable 5, has captured only about 11% of total product spending across its lineup more than two months after launch, with growth nearly stalling. Remarkably, its own lower-priced model, Opus 5, has already overtaken Fable 5 in enterprise spending. Enterprises are increasingly shifting toward cheaper models with 'good enough' performance, prompting a reevaluation of the business models of top AI firms that heavily invest in cutting-edge models.

According to the Financial Times, payment platform Ramp, which tracks spending across 70,000 companies, found that Fable 5 accounted for only about 11% of total spending on Anthropic products since its launch in early June. Growth momentum has recently plateaued.

Analysts and Anthropic investors point to Fable 5's high price as a key factor, along with the fact that older, cheaper models are already sufficient for most enterprise tasks.

More notably, Anthropic's own lower-priced, smaller-scale Opus 5, launched at the end of July, has already surpassed Fable 5 in enterprise spending—indicating a shift in enterprise AI adoption from pure performance pursuit to a balance of cost and practical needs.

Competitors are adding further pressure. OpenAI's GPT-5.6 is priced significantly below Fable 5, and its parent company has seen its annualized revenue grow 35% this quarter, surpassing $400 billion. Meanwhile, the growing number of lower-cost open-source models gives enterprises more options, further squeezing the pricing power of top U.S. AI labs.

Fable 5 also faced policy-related setbacks. In early June, the Trump administration temporarily forced Anthropic to withdraw the model over national security concerns, impacting the company's revenue growth that month. The model was allowed to relaunch on July 1.

However, analysts and investors note that with restrictions lifted, political uncertainty is no longer a major factor in enterprise model selection. Price and performance have regained prominence as key decision drivers.

Miles Clements, partner at Accel, stated that most enterprises don't actually need to run on the absolute frontier of model capability, and the era of enterprises blindly choosing the strongest model 'was not a sustainable phase.'

This shift is particularly sensitive for Anthropic, as frontier AI labs have long invested billions in developing larger, more complex models, expecting to recoup costs through premium pricing.

If enterprises continue shifting toward cheaper, 'good enough' performance models, the logic of 'driving revenue through the strongest models' could face fundamental challenges.

Nonetheless, Anthropic's overall business remains on a high-growth trajectory. Revenue has increased nearly sevenfold since the beginning of the year. The company achieved adjusted operating profitability for the first time in Q2 and has signaled to investors it may be profitable again in Q3. The number of customers with annual spending exceeding $100,000 has reached 6,000.

Anthropic's annualized revenue in July reached $65 billion, up from $47 billion in May, though below some investors' initial expectations of over $80 billion.

As Anthropic prepares for an IPO, with market expectations of a valuation exceeding $2 trillion, the adoption rate of Fable 5 has become a key market indicator.

Ara Kharazian, chief economist at Ramp, admitted that predicting Anthropic's trajectory over the coming months is 'almost an impossible task.' As enterprises increasingly prioritize cost-effectiveness, AI companies face new challenges in sustaining revenue growth through ever-stronger, more expensive models.