Claude Fable 5.1: Anthropic Cuts Cache-Read Prices 75%
By the ChatGBot Team · Published October 12, 2026
Anthropic has shipped Claude Fable 5.1 with one headline change: the price of a cache read dropped 75%, from $1 to $0.25 per million tokens. It sounds like a small line item, but for long-running agents that reuse the same context again and again, it meaningfully lowers the bill. Here is what changed and who it helps.
Every top AI in one app
GPT, Claude, DeepSeek, Qwen and GLM — one subscription.
Try ChatGBot →What changed in Fable 5.1
The one line on the rate card that moved is the cache read. A cached input token now costs $0.25 per million, down from $1 on Fable 5, a 75% cut. Base input and output prices held steady, reported at around $10 and $50 per million tokens. So this is a targeted price change aimed at a specific pattern, not an across-the-board discount.
What a cache read actually is
When an app sends the same context to the model repeatedly, such as a long system prompt, a document, or a growing chat history, that context can be cached and reused instead of reprocessed from scratch each time. Reusing it is a cache read, and it is now four times cheaper. The more your workload repeats the same context, the more you save.
Who benefits most
Agents. Long-running, multi-step AI agents loop over the same instructions and context constantly, so they lean heavily on cache reads. Anthropic's own estimates, as reported, put the saving at roughly 25% on typical workloads and up to about 45% on highly agentic ones. If your usage rarely repeats context, the benefit is smaller.
Fable 5.1 vs Mythos 5.1
Alongside Fable 5.1, Anthropic also referenced Mythos 5.1. Reporting describes them as the same underlying model shipped with different safeguards: Fable 5.1 is generally available, while Mythos 5.1 is restricted to vetted participants in trusted-access programs aimed at cybersecurity defenders and life-sciences researchers.
How it fits the Claude lineup
This pricing move sits within a busy Claude 5 era that also brought Claude Opus 5.5 and Sonnet 5.5. For how the family fits together, see Claude models explained. The pattern across all of them is clear: the frontier keeps getting cheaper to run, especially for agent-style work.
Frequently asked questions
What is new in Claude Fable 5.1? The cache-read price fell 75%, from $1 to $0.25 per million tokens. Base input and output prices stayed the same.
How much can I actually save? The 75% cut applies only to cache reads, so total savings depend on how much your workload reuses context. Reported estimates are about 25% on typical use and up to roughly 45% on heavily agentic use.
What is the difference between Fable 5.1 and Mythos 5.1? They are reported to be the same model with different safeguards. Fable 5.1 is generally available, while Mythos 5.1 is restricted to vetted trusted-access participants.
Can I use Claude in ChatGBot? Yes. ChatGBot gives you Claude alongside GPT, DeepSeek, Qwen and GLM under one subscription, so you can switch models mid chat.
Use Claude and more models in one app
With ChatGBot you get Claude together with GPT, DeepSeek, Qwen and GLM under a single subscription, so you can pick the right model for each task without paying for several tools.