🤖Anthropic Launches Claude Fable 5 With Safety Limits
Fable 5: The Most Powerful AI with Built-In Safeguards
TL;DR
Anthropic launches Claude Fable 5, a powerful AI model with strict safety limits. It excels in software engineering and vision but blocks high-risk responses, switching to Opus 4.8 when needed. Access is expanding globally, with pricing set at $10 per million input tokens.
Anthropic just dropped Claude Fable 5, their most powerful AI model yet, complete with safety limits in place for cybersecurity and other risky areas. If you're working on critical systems or need advanced reasoning skills from an AI, this is a big deal. The model blocks unsafe responses and falls back to Opus 4.8 when necessary, ensuring safer operations. Fable 5 costs $10 per million input tokens and $50 for output tokens, with no extra cost until June 23rd for existing plans. But starting June 24th, you'll need usage credits, making it a bit pricier.
Key Points
Claude Fable 5 excels in software engineering, knowledge work, and vision tasks; $10 per million input tokens
$50 per million output tokens for Fable 5; no extra cost until June 23rd for existing plans
Fable 5 blocks responses in high-risk areas like cybersecurity, falling back to Claude Opus 4.8 when needed
Over 1,000 hours of testing produced no universal jailbreaks, with external red-teaming also failing to find them
Access on subscriptions will roll out in stages: Fable 5 is included in Pro, Max, Team, and seat-based Enterprise plans until June 23rd
Why It Matters
If you're working on critical infrastructure or need advanced reasoning skills from an AI model like Claude Fable 5, this launch could be a game changer. The safety limits ensure safer operations in high-risk areas, but the $10 per million input tokens pricing might make smaller teams think twice.
Frequently Asked Questions
Why does this matter?
If you're working on critical infrastructure or need advanced reasoning skills from an AI model like Claude Fable 5, this launch could be a game changer. The safety limits ensure safer operations in high-risk areas, but the $10 per million input tokens pricing might make smaller teams think twice.
What happened?
Anthropic launches Claude Fable 5, a powerful AI model with strict safety limits. It excels in software engineering and vision but blocks high-risk responses, switching to Opus 4.8 when needed. Access is expanding globally, with pricing set at $10 per million input tokens.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.