Anthropic Launches Claude 4 Models With Advanced Coding and Agent Capabilities
Anthropic unveils Claude Opus 4 and Claude Sonnet 4, delivering state-of-the-art coding performance and new hybrid reasoning modes. The update also brings the general availability of Claude Code and powerful new API tools for developers.
Anthropic introduces Claude Opus 4 and Claude Sonnet 4 as its next generation of AI models, setting new standards for coding and advanced reasoning. Claude Opus 4 stands as the world's best coding model by leading major benchmarks like SWE-bench, while Claude Sonnet 4 delivers significant upgrades over its predecessor in both reasoning and instruction following. Both models operate as hybrid systems that offer near-instant responses alongside an extended thinking mode for deeper problem solving.
The new models feature extended thinking that integrates tool use, allowing the AI to alternate between reasoning and actions like web search to produce better answers. Claude also gains the ability to use multiple tools in parallel, follow instructions with greater precision, and build long-term memory by extracting and saving key facts from local files. Additionally, Claude Code exits its research preview to become generally available, bringing native integrations to VS Code and JetBrains alongside support for background tasks through GitHub Actions.
Developers receive four new API capabilities to build more powerful AI agents, including a code execution tool, an MCP connector, a Files API, and prompt caching for up to one hour. The models are accessible across Claude's Pro, Max, Team, and Enterprise plans, with Sonnet 4 also available to free users. Both models launch on the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI, maintaining the same pricing structure as previous generations at $15/$75 per million tokens for Opus 4 and $3/$15 for Sonnet 4.