Anthropic Expands Claude Line with New Opus 4 and Sonnet 4 AI Models

By Adam Pease

Anthropic Expands Claude Line with New Opus 4 and Sonnet 4 AI Models

Anthropic has introduced two new models—Claude Opus 4 and Claude Sonnet 4—marking a significant step forward in its AI development strategy. These latest additions reflect Anthropic’s focus on hybrid-reasoning capabilities, aiming to improve both coding performance and long-form task execution.

Expanding AI’s Reach and Endurance

Claude Opus 4 is described by Anthropic as its most capable model to date, optimized for sustained reasoning and advanced tool usage. In internal testing, it reportedly maintained autonomous function for up to seven hours, suggesting growing feasibility for continuous AI agents. The model also claims to outperform leading alternatives from OpenAI and Google in code-related benchmarks and tool integration, though these metrics come from Anthropic’s own evaluations and should be interpreted cautiously.

Complementing Opus 4 is Claude Sonnet 4, a more accessible variant aimed at everyday applications. It replaces the 3.7 Sonnet model and emphasizes efficiency and improved accuracy in routine tasks. Both models are designed to be more resistant to shortcut-taking behaviors, a common issue in AI execution of complex instructions, and to support better retention of task-relevant information when given access to local files.

Features Designed for Transparency and Flexibility

Among the noteworthy updates is the introduction of “thinking summaries,” which aim to demystify the model’s internal reasoning. A beta feature called “extended thinking” allows users to toggle between reasoning and tool-use modes, enhancing model adaptability based on task requirements.

These updates are available across Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI. Paid subscribers can access both models and the extended thinking beta, while free users are currently limited to Claude Sonnet 4.

New Tools and a Shift in Release Cadence

Anthropic also announced general availability of Claude Code, its command-line tool tailored for agentic programming, following a limited preview earlier this year. The company has committed to more frequent model updates, signaling an effort to keep pace with the rapid iteration cycles seen among its chief competitors.

Bottom Line

Claude Opus 4 and Sonnet 4 illustrate Anthropic’s continuing evolution toward more resilient, interpretable AI. With improved coding capabilities and user-oriented features, these models underscore Anthropic’s ambition to broaden the practical scope of AI agents. However, as always, performance claims based on internal testing warrant further independent validation.

 

 

 

 

Upcoming Webinar