Anthropic releases Claude Fable 5.1 with 75% cheaper cache reads, cutting costs up to 45%
Anthropic released Claude Fable 5.1 and Mythos 5.1, the same underlying model with different safety restrictions. Fable 5.1 is available to all API customers, while Mythos 5.1 is limited to vetted researchers. The models feature a 1 million token context window, improved agentic coding and research capabilities, and significant cost reductions: cache reads dropped 75% to $0.25 per million tokens, reducing overall costs by up to 45% for highly agentic tasks. Performance benchmarks show major gains over previous generations.
Editorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page itself is projected from evidence records.
- Current automated evidence projection
Cross-source coverage
Wire timeline
Fable 5.1 maintains Fable 5 pricing at $10/M input and $50/M output tokens
Fable 5.1, the latest iteration of the Fable AI model series, retains the same pricing structure as its predecessor Fable 5. Input tokens remain priced at $10 per million tokens, while output tokens stay at $50 per million tokens. A significant price reduction has been applied to prompt cache reads, which now cost $0.25 per million tokens, representing a 75% decrease from the previous cache-read rate. This pricing update was shared via an OpenRouter post, indicating the model is available through that platform. The unchanged pricing for standard tokens and the steep discount for cached reads aim to provide cost predictability and encourage efficient use of cached prompts.
Anthropic slashes Fable 5.1 cache read price by 75% to $0.25 per million tokens
Anthropic has implemented a targeted price cut for its Fable 5.1 AI model, focusing on reducing cache read costs rather than standard input/output prices. While input and output prices remain unchanged at $10 and $50 per million tokens respectively, the cache read price has been slashed by 75% to just $0.25 per million tokens. This results in approximately 25% savings for regular tasks, but agent tasks that repeatedly read context can see savings of up to 45%. The pricing strategy is clearly aimed at long-running AI agents that benefit most from reduced cache read costs, making them more economical for extended operations.
Anthropic releases Claude Fable 5.1 and Mythos 5.1 with 1M token context and lower cache costs
Anthropic has released Claude Fable 5.1 and Mythos 5.1 via its API, according to official documentation. The new models feature a 1 million token context window with 128K maximum output, improved capabilities for agentic coding, research, document handling, spreadsheets, and slides, as well as better computer use and long-horizon task performance. They include adaptive thinking with adjustable effort. Pricing is set at $10 per million input tokens and $50 per million output tokens, with cached input tokens at $0.25 per million—75% cheaper cache reads than the previous Fable 5 model. Fable 5.1 is available to all API customers, while Mythos 5.1 is currently limited to participants in Project Glasswing.
Show 3 older updatesHide older updates
Anthropic releases Claude Fable 5.1 and Mythos 5.1 with major performance gains
Anthropic has released Claude Fable 5.1 and Claude Mythos 5.1, which are the same underlying model differentiated by safety restrictions. Fable 5.1 is available to all users, while Mythos 5.1 is restricted to vetted cybersecurity and life sciences researchers. Performance benchmarks show significant improvements: Fable 5.1 scored 52.6% on Terminal-Bench-Science, up from 24.7% in the previous generation, surpassing Opus 5's 29.0%. On Terminal-Bench 4.0, Fable 5.1 scored 55.8%, Mythos 5.1 reached 60.9%, and Opus 5 scored 52.3%. The model demonstrated autonomous capabilities, including running for 38 consecutive hours in a Ramp test and tracing a rare crash to its root cause in a MongoDB test. Mythos 5.1 achieved a 50% experimental hit rate in protein binder design and optimized GPU kernels for biological models. Pricing has been reduced, with Cache Read dropping 75% to $0.25 per million tokens, and overall costs decreasing by up to 45% for highly agentic tasks.
Fable 5.1 cache reads cost 75% less than Fable 5, reducing model costs by up to 45%
A post from ClaudeAI announces that cache reads with Fable 5.1 cost 75% less than with Fable 5. This improvement reduces the overall cost of the model in practice by approximately 25% for typical workloads and up to 45% for highly agentic ones. The announcement highlights significant cost efficiency gains for users of the updated AI model, particularly for tasks involving extensive agentic behavior.
Unnamed AI model claims higher performance than Fable 5 at lower cost
A post on X, attributed to a Claude AI source, announces an unnamed AI model that claims to outperform Fable 5 in performance while also achieving similar or better results at a significantly lower cost when set to lower effort levels. The post includes a link to further details. The comparison suggests a new AI model offering competitive advantages in both capability and cost efficiency, potentially impacting the AI model market.