Anthropic has reduced the price of a Fable 5.1 cache hit to $0.25 for input, down from $1 for Fable 5. The 75% cut could lower costs for developers whose applications repeatedly send the same stored context to a model.
The new cache-hit price equals 2.5% of Fable’s standard $10 input-token price. By comparison, most other Claude models use a cache-hit rate equal to 10% of their normal input price.
How the New Pricing Compares
Prompt caching lets an application reuse previously processed input. That input may include long instructions, documents, code, or conversation history. A cache hit occurs when the model can retrieve that stored material instead of processing it again at the regular rate.
The announced figures produce three key comparisons:
- Fable 5.1 cache-hit input costs $0.25, compared with $1 for Fable 5.
- The reduction from the prior version is 75%.
- The cache-hit rate is 2.5% of the normal $10 input price, rather than 10%.
“Anthropic has cut a Fable 5.1 cache hit to just $0.25 on input, down from $1.00 for Fable 5.”
The comparison indicates that Fable 5’s earlier $1 cache-hit price matched the 10% pricing pattern used by most other Claude models. Fable 5.1 breaks from that pattern with a deeper discount.
Why Cache Costs Matter
Input charges can form a large share of spending for applications built around long prompts. Customer-support systems may reuse policy documents. Coding tools may repeatedly submit repositories or technical instructions. Research services may rely on the same source files across many requests.
For those workloads, a lower cache-hit price can reduce the cost of repeated context. The largest benefit would go to services with high request volumes and strong cache-hit rates.
The savings may be smaller for applications that use short prompts or frequently changing input. A request that does not match stored content may be billed under standard input pricing. Output-token charges and any cache-writing fees may also affect the final bill.
Developers Still Need Full Cost Tests
The headline discount does not reveal the total cost of running a Fable 5.1 application. The available figures cover cache-hit input pricing, not every charge connected to a request.
Teams assessing the change should measure how often cached material is reused. They should also compare response quality, speed, output use, and the cost of creating or refreshing cache entries.
Even with those limits, the reduction changes the economics of repeated-input workloads. At $0.25, a Fable 5.1 cache hit costs one-quarter of the Fable 5 rate and one-fortieth of normal Fable input.
The next issue to watch is whether Anthropic applies similarly low cache ratios to other Claude models. For now, Fable 5.1 stands apart from the company’s more common 10% multiplier and gives high-reuse applications a stronger reason to examine caching.
Rashan is a seasoned technology journalist and visionary leader serving as the Editor-in-Chief of DevX.com, a leading online publication focused on software development, programming languages, and emerging technologies. With his deep expertise in the tech industry and her passion for empowering developers, Rashan has transformed DevX.com into a vibrant hub of knowledge and innovation. Reach out to Rashan at [email protected]
