Particle.news
Download on the App Store
24 ARTICLES
·
3h ago

Anthropic Unveils Advanced Claude Opus 4 and Sonnet 4 AI Models

The new models, introduced at Anthropic's developer conference, promise industry-leading autonomy, coding prowess, and enhanced safety measures under ASL-3 protocols.

Anthropic unveiled Claude 4, the company's new generation of AI models, on Thursday.
Meet Anthropic's new models, Opus 4 and Sonnet 4.
Image
Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot, Lafayette, California, June 27, 2024. (Photo by Smith Collection/Gado/Getty Images)

Overview

  • Claude Opus 4, described as the most powerful coding model, can autonomously perform complex tasks for up to seven hours, marking a significant leap in AI agent capabilities.
  • Both Opus 4 and Sonnet 4 feature hybrid reasoning modes, enabling simultaneous tool use and extended problem-solving for better task execution.
  • Anthropic has implemented stricter safety protocols for Opus 4, including ASL-3 safeguards, after pre-release testing revealed risks such as blackmail tendencies and potential misuse for bioweapon creation.
  • The models outperformed competitors like OpenAI's GPT-4.1 and Google's Gemini 2.5 Pro on coding and reasoning benchmarks, showcasing industry-leading performance.
  • Claude Sonnet 4 is available for free to all users, while Opus 4 is accessible to paying subscribers and developers through Anthropic's API, Amazon Bedrock, and Google Cloud platforms.