Claude Fable 5.1 Explained: Price, Cache and Specs
In short: Claude Fable 5.1 is an Anthropic AI model released on 1 September 2026, the successor to Claude Fable 5. It costs $10 per million input tokens and $50 per million output tokens - the same as Fable 5 - with a 1M-token context window and up to 128k tokens of output. The one pricing change is the prompt cache: a cache read costs $0.25 per million tokens, 2.5% of the input price, against $1.00 on Fable 5.
What is Claude Fable 5.1?
Claude Fable 5.1 is a large language model from Anthropic, available in the Claude apps and through the Claude API. It reads text and images and writes text. Anthropic's models overview lists it among its current models, with Claude Fable 5 moved to the legacy list. Its API model ID is claude-fable-5-1.
For what a large language model is and how one is trained, see how LLMs work. The older model is covered in Claude Fable 5 explained.
How much does Claude Fable 5.1 cost?
Input costs $10 per million tokens and output $50, the same as Fable 5. The price list from Anthropic's pricing page:
| Item | Claude Fable 5.1 | Claude Fable 5 |
|---|---|---|
| Input / 1M tokens | $10 | $10 |
| Output / 1M tokens | $50 | $50 |
| Cache read / 1M tokens | $0.25 (2.5% of input) | $1.00 |
| Batch API | 50% off | 50% off |
A token is a small chunk of text - see tokens and context windows explained.
What does the cheaper cache read change?
It lowers the cost of input the model has already seen, and nothing else. Prompt caching stores a long, repeated part of a prompt - a system prompt, a codebase, a set of documents - so later requests that reuse it are billed at the cache-read rate instead of the full input rate. On Fable 5.1 that rate fell from $1.00 to $0.25 per million tokens.
Requests that reuse a large cached prefix on every call see the difference; requests that are short or all different barely use the cache. Input and output tokens cost the same as on Fable 5.
What are Claude Fable 5.1's specs?
| Spec | Claude Fable 5.1 |
|---|---|
| Developer | Anthropic |
| Released | September 1, 2026 |
| Context window | 1M tokens |
| Maximum output | 128k tokens |
| Input | Text and images |
| Output | Text |
| Reasoning | Adaptive thinking, always on; default effort high |
| Knowledge cutoff | June 2026 |
| Licence | Proprietary - not downloadable |
What are effort levels on Claude Fable 5.1?
Effort is a setting that controls how much the model thinks before answering; it is one model, not several. Higher effort spends more thinking tokens - billed as output - before the reply. Anthropic's default is high. Leaderboards that test several settings list the same model more than once, one row per effort level. Adaptive thinking is always on and cannot be switched off, as with Fable 5.
Is Claude Fable 5 still available?
Yes. Anthropic lists Claude Fable 5 under legacy models. In Anthropic's documentation, legacy means still callable, with retirement dates published in advance - not withdrawn. Moving from Fable 5 to Fable 5.1 in the API is a change of model ID, to claude-fable-5-1.
How does Claude Fable 5.1 rank?
On LMArena's text leaderboard Claude Fable 5.1 has 1504 Elo (rank 5, max effort). The full ranked table, with every model's effort settings, is on our benchmarks page.
Words to know
- Token - a chunk of text a model reads or writes; prices are per million.
- Context window - the most tokens a model can handle in one request.
- Prompt caching - storing repeated input so later requests reuse it at a lower rate.
- Cache read - input billed from the cache instead of at the full rate.
- Effort level - how much a model reasons before answering.
- Legacy model - an older model still available, with a retirement date announced in advance.
Related guides
Frequently asked questions
What is Claude Fable 5.1?+
Claude Fable 5.1 is a large language model from Anthropic, released on 1 September 2026, that reads text and images and writes text. It succeeds Claude Fable 5 and is available in the Claude apps and the Claude API as claude-fable-5-1.
How much does Claude Fable 5.1 cost?+
It costs $10 per million input tokens and $50 per million output tokens, the same as Claude Fable 5. Cache reads cost $0.25 per million tokens, and the Batch API is 50% off.
What is the difference between Claude Fable 5 and Fable 5.1?+
Token prices, the 1M-token context window and the 128k maximum output are the same. The documented price change is the cache read: $0.25 per million tokens on Fable 5.1, against $1.00 on Fable 5.
What is Claude Fable 5.1's context window?+
Claude Fable 5.1 has a context window of 1M tokens and can write up to 128k tokens in one reply.
Can you turn off thinking on Claude Fable 5.1?+
No. Adaptive thinking is always on, as with Claude Fable 5. The effort setting controls how much it thinks, with high as Anthropic's default.
Enjoyed this? Get more in your inbox.
Up to ten of the week's AI news stories, by email once a week.