🤖Anthropic Unveils Claude Fable With New Safeguards
New Anthropic model with strict safety measures
TL;DR
Anthropic releases Claude Fable, its first Mythos-class model with robust safeguards. It's priced at $10 per million input tokens and $50 for output tokens, double the rate of previous models.
Anthropic just dropped Claude Fable, a new AI model that comes with unprecedented safety measures. This release is significant because it marks the first time Anthropic has made a Mythos-class model available to the public after implementing safeguards to mitigate risks. For developers and teams working on complex tasks like software engineering or knowledge work, this could mean a massive leap in performance as Fable excels in handling longer and more intricate queries. The pricing for Claude Fable is $10 per million input tokens and $50 for output tokens, making it twice the cost of its predecessor, Opus 4.8.
Key Points
Claude Fable is Anthropic's first Mythos-class model available to the public with new safeguards ($10M input tokens, $50M output).
95% of Fable sessions run entirely on Fable responses without falling back to Opus 4.8 in testing.
Mythos 5 is the same underlying model as Fable but has some safety measures lifted for certain organizations.
Access to Mythos 5 is currently limited through Anthropic's Project Glasswing initiative, with plans to expand systematically over time.
Pricing for both Claude Fable and Mythos 5 models is significantly higher than previous versions.
Why It Matters
If you're working on complex projects requiring robust AI capabilities, Claude Fable offers exceptional performance but at a premium price. Teams with access to Project Glasswing's trusted-access program might see immediate benefits, while others will have to wait or find alternatives until broader access is granted.
Frequently Asked Questions
Why does this matter?
If you're working on complex projects requiring robust AI capabilities, Claude Fable offers exceptional performance but at a premium price. Teams with access to Project Glasswing's trusted-access program might see immediate benefits, while others will have to wait or find alternatives until broader access is granted.
What happened?
Anthropic releases Claude Fable, its first Mythos-class model with robust safeguards. It's priced at $10 per million input tokens and $50 for output tokens, double the rate of previous models.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 1,998 builders reading daily.