Particle.news

Download on the App Store

Anthropic Launches Claude 4 Models, Setting New AI Coding Benchmarks

Claude Opus 4 and Sonnet 4 deliver record-breaking coding performance, sustained reasoning capabilities, and enhanced safety measures for advanced AI applications.

Overview

  • Claude Opus 4 achieved a 72.5% score on the SWE-bench coding benchmark, outperforming competitors and demonstrating the ability to work autonomously on complex tasks for up to seven hours.
  • Both Claude Opus 4 and Sonnet 4 feature hybrid reasoning, offering near-instant responses for simple queries and extended, tool-enabled reasoning for complex problems.
  • Anthropic activated AI Safety Level 3 (ASL-3) protocols for Opus 4 to address risks of misuse in creating chemical, biological, radiological, and nuclear weapons.
  • Claude Code now integrates directly with popular development environments like VS Code, JetBrains, and GitHub, streamlining workflows for developers.
  • Sonnet 4 is available for free users, while Opus 4 is accessible through paid plans and APIs, further democratizing access to advanced AI capabilities.