Ramp Finds Fable 5 Drew 75% of GPT-5.6 Sol Spend
Ramp reported on August 12 that Claude Fable 5 generated about 75% as much model-attributed business spend as OpenAI's GPT-5.6 Sol in July. Fable accounted for 6% of Anthropic tokens and 11.4% of its model spend in Ramp's sample, while Sol represented 25% of OpenAI tokens and 23% of spend.
Ramp reported on August 12 that Anthropic's Claude Fable 5 generated approximately 75% as much model-attributed spend as OpenAI's GPT-5.6 Sol during July. The finding comes from Ramp's token spend management product, which tracks daily model usage and cost data for participating businesses.
The comparison does not mean OpenAI led Anthropic across business adoption. Ramp's broader AI Index found that 43.5% of U.S. businesses in its sample paid for Anthropic subscriptions or tokens in July, compared with 39.7% for OpenAI. Instead, the model-level data points to different uptake of the two companies' highest-end offerings.
Fable represented a smaller share of Anthropic usage
Ramp said Fable 5 accounted for 6% of tokens purchased from Anthropic and 11.4% of dollars spent on Anthropic models in the sample. GPT-5.6 Sol represented 25% of OpenAI tokens and 23% of OpenAI model spend.
Ramp lead economist Ara Kharazian interpreted the gap as evidence that some businesses are resisting the price premium for the most capable model. The company lists Fable 5 at roughly $10 per million input tokens, about twice the input price of GPT-5.6 Sol. That interpretation should remain bounded by the sample: Ramp says the model-level group skews more technical than its broader AI Index population, and the figures measure spending among businesses connected to Ramp's product rather than the entire market.
Business Insider independently reported the same July comparison and noted that Fable's month included a temporary government-ordered withdrawal, a July 1 return, and pricing changes. Those disruptions complicate any attempt to treat one month's spending as a stable verdict on product demand.
Spending is still concentrated at the top
Ramp's first-party report also shows how uneven business AI spending remains. In July, the top 1% of businesses in its sample spent a median $7,400 per employee on AI, while the median firm spent $11.95 per employee. Ramp said the top 10% spent a median $650 per employee.
For teams selecting models, the practical signal is about willingness to pay, not a universal capability ranking. A model can lead selected evaluations yet capture a smaller share of enterprise workloads if the incremental quality does not justify its cost for common tasks. The July figures make routing and workload-level evaluation more important: organizations need to test where premium reasoning changes outcomes enough to offset higher token costs, while avoiding broad conclusions from a single vendor sample and a single launch month.
Key Points
- 1Ramp says Fable 5 generated about 75% as much model-attributed spend as GPT-5.6 Sol in its July business sample.
- 2Fable represented 6% of Anthropic tokens and 11.4% of its model spend; Sol represented 25% of OpenAI tokens and 23% of spend.
- 3Ramp's sample and Fable's disrupted launch month limit how broadly the comparison can be generalized.
Scoring Rationale
First-party spending data provides a concrete signal about business uptake and price sensitivity for frontier models. The comparison is useful for model-routing and procurement decisions, but its significance is bounded by Ramp's participating-business sample and Fable 5's disrupted first month.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
