Background
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on 1 September 2026. [2][3]
What happened
Anthropic called Fable 5.1 the world's most advanced model for coding and knowledge work. [1]
Prices stayed at $10 per million input tokens and $50 per million output tokens. Cache reads were cut by 75%, and Anthropic estimated savings of up to about 45% on agentic work. [1]
Key details
Fable 5.1 scored 55.8% on Terminal-Bench 4.0, against 42.0% for Fable 5. [1]
Mythos 5.1 became available through the Cyber Verification Program and a new Life Sciences Verification Program. [1]
Anthropic said models released after 2 August 2026 include an invisible text watermark. [1]
A cache read is a lower-cost charge for text the model has already processed, which is why a cut to that price matters most for long, repeated inputs. Terminal-Bench is a benchmark that tests whether a model can complete tasks in a command-line terminal. A verification programme gives access to a sensitive capability only after the organisation has been checked.
Impact
Fable 5.1 is generally available, while Mythos 5.1 is offered through verification programmes. [1][2][3]
Sources
- 1
- 2Model deprecationsAnthropic (Claude Platform Docs)
- 3Claude (language model)Wikipedia
Cite this page: Anthropic releases Claude Fable 5.1 and Mythos 5.1. TellWorld, https://www.tellworld.net/news/claude-fable-5-1-2026. Updated 2026-11-08