7 veículos·13 reportagens
Anthropic recua política que limitava a capacidade do Claude Fable 5 de desenvolver outros modelos de IA, após forte reação da comunidade de pesquisa em IA (Maxwell Zeff/Wired)
Anthropic lançou o Claude Fable 5, um modelo de classe Mythos agora disponível ao público. O modelo custa US$ 10 por milhão de tokens de entrada e US$ 50 por milhão de tokens de saída. Ele utiliza classificadores de segurança que acionam um fallback para o Claude Opus 4.8 em cerca de 5% das sessões e bloqueiam a discussão de determinados tópicos perigosos.
Escrito a partir de todas as reportagens abaixo, não de uma só.
Tradução automática. As reportagens originais estão no idioma de origem.
Como foi noticiado
- The Verge·Anthropic releases its first Mythos-class model Claude Fable
- TechMeme·Anthropic releases Claude Fable 5, a "safe" Mythos-class model it says can't be used for cyberattacks, to the public, and Claude Mythos 5 to trusted orgs (Wired)
- TechMeme·Anthropic prices both Claude Fable 5 and Mythos 5 at $10 per 1M input and $50 per 1M output tokens, less than half the price of Claude Mythos Preview (David Gewirtz/ZDNET)
- VentureBeat·Anthropic brings Mythos to the masses with Claude Fable 5, its most powerful generally available model ever
- TechMeme·Anthropic says Fable 5 uses conservative safety classifiers that trigger a fallback to Claude Opus 4.8 in ~5% of sessions, in areas like cybersecurity (Anthropic)
- The Next Web·Anthropic releases Claude Fable 5, a Mythos-class model the public can finally use, days before a potential record IPO
- Ars Technica·Anthropic says these topics are too dangerous to let its Fable 5 model talk about
- Engadget·Anthropic's Fable AI brings the capabilities of its unreleased Mythos model to regular users
- The Verge·Microsoft AI head calls out Anthropic for acting like Claude is conscious
- TechMeme·Anthropic says Fable 5 has invisible safeguards that use prompt modification, steering vectors, or PEFT to limit its effectiveness for frontier LLM development (Matthias Bastian/The Decoder)
- TechCrunch·Cybersecurity researchers aren’t happy about the guardrails on Anthropic’s Fable
- TechMeme·Anthropic secretly limiting Claude's usefulness for LLM development strengthens the argument that Anthropic is using AI safety to justify monopolistic behavior (Dean W. Ball/@deanwball)
- TechMeme·Anthropic backtracks on a policy limiting Claude Fable 5's ability to develop other AI models, after significant backlash from the AI research community (Maxwell Zeff/Wired)