Anthropic launches Claude 4

Anthropic has unveiled its latest AI models, Claude Opus 4 and Claude Sonnet 4, which set a new benchmark in coding, complex reasoning, and AI agent performance. Claude Opus 4 is hailed as the world’s top coding model, excelling in long-running, complex tasks with standout scores on SWE-bench (72.5%) and Terminal-bench (43.2%). Claude Sonnet 4, a significant upgrade from Sonnet 3.7, combines performance and efficiency with improved steerability and instruction-following.

A standout feature of both models is extended thinking with tool use, allowing the AI to alternate between reasoning and tool use, improving the depth and accuracy of its responses. These models can also execute tools in parallel and demonstrate improved memory when given access to local files, enabling better long-term task tracking.

Claude Code, now generally available, extends the models into software development workflows. Integrated with VS Code and JetBrains, Claude offers in-line edits and background support via GitHub Actions. A new Claude Code SDK also lets developers build custom agents with the same core technology.

The models are accessible via the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. Pricing remains unchanged: Opus 4 at $15/$75 and Sonnet 4 at $3/$15 per million input/output tokens. Sonnet 4 is available even to free users, while both models are included in Pro, Max, Team, and Enterprise plans.

Anthropic’s launch marks a major leap toward creating virtual collaborators that maintain context, sustain long-term focus, and enable powerful new workflows for developers and enterprises alike.