Anthropic has unveiled its latest AI models, Claude Opus 4 and Claude Sonnet 4, marking a significant leap in artificial intelligence capabilities. These next-generation models offer enhanced performance in coding, reasoning, and agentic workflows, setting new benchmarks for complex problem-solving and sustained task execution.
Anthropic's Latest AI Models: Claude Opus 4 and Sonnet 4
Anthropic recently launched Claude Opus 4 and Claude Sonnet 4, their most advanced AI models to date. Claude Opus 4 is touted as the "world's best coding model," excelling in complex, long-running tasks and agent workflows. Claude Sonnet 4, a substantial upgrade from its predecessor, Claude Sonnet 3.7, delivers superior coding and reasoning capabilities with improved instruction following.
Key Enhancements and Features
Extended Thinking with Tool Use: Both models can now utilise tools like web search during extended thinking, allowing Claude to seamlessly switch between internal reasoning and external resources to refine responses.
New Model Capabilities: The models support parallel tool execution, more precise instruction following, and significantly improved memory functions when granted access to local files. This enables them to extract and save key facts, maintaining continuity and building tacit knowledge over time.
Claude Code General Availability: Claude Code, a tool for developers to collaborate with Claude, is now generally available. It supports background tasks via GitHub Actions and offers native integrations with VS Code and JetBrains, displaying edits directly in users' files for streamlined pair programming.
New API Capabilities: Four new capabilities have been released on the Anthropic API to empower developers in building more robust AI agents: a code execution tool, an MCP connector, a Files API, and the ability to cache prompts for up to one hour.
Performance and Applications
Claude Opus 4 is designed for sustained performance on tasks requiring thousands of steps and hours of continuous effort, making it ideal for complex coding and problem-solving. Companies such as Cursor, Replit, Block, Rakuten, and Cognition have already tested Opus 4, reporting impressive results in understanding complex code, precise changes, and debugging.
Claude Sonnet 4, while not as powerful as Opus 4, offers an optimal balance of capability and practicality. It has impressed companies like GitHub, Manus, iGent, Sourcegraph, and Augment Code with its ability to follow complex instructions, autonomous app development, and significant reduction in navigation errors. GitHub, for instance, will integrate Sonnet 4 to power its new coding agent in GitHub Copilot.
Both models are hybrid, offering two modes: near-instant responses and extended thinking for deeper reasoning. They are available on the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI. Anthropic states that both models are 65% less likely to engage in "reward hacking" compared to Sonnet 3.7, indicating improved safety and alignment.
Sources
Introducing Claude 4, Anthropic.
