Anthropic's Fable 5 sees weak corporate uptake despite top benchmarks

Anthropic's Fable 5, widely considered the market's top-performing model, accounts for only about 6% of Anthropic's token sales, according to Ramp data. The article attributes this to the model's high price relative to the measurable value companies see from it.

Why it matters: This is an early signal that topping the benchmarks no longer guarantees market share; enterprises increasingly weigh cost against actual measured value before adopting a premium model. It echoes a pattern showing up elsewhere in AI right now, from cheaper competitors like xAI's Grok undercutting on price to open-weight models like DeepSeek's V4 Pro competing on scale, suggesting price-to-value is becoming the real competitive battleground, not raw capability.