May 23rd, 2011 - Anthropic, Inc. today (May 23rd) at 0:30pm BST, launched the Claude Opus 4 and Claude Sonnet 4 Next Generation Language Models.Achieve significant advances in the areas of structured reasoning, software engineering, and autonomous agent behavior.
Claude Opus 4: The pinnacle of complex reasoning and software development
Claude Opus 4 is positioned as Anthropic's most powerful model to date, designed to handle complex reasoning processes and software development scenarios, 1AI cites a blog post.
The test data shows that the model achieves an accuracy of 72.51 TP3T in the SWE-bench benchmark test (which evaluates the model's ability to solve real GitHub problems), and 43.21 TP3T in the TerminalBench test (which validates the model's performance in a multi-step terminal code generation task).
Even more strikingly, Opus 4 exhibits strong autonomous behavior in software environments, thanks to improved memory management, broader context retention, and stronger internal planning mechanisms, according to Rakuten test data thatNearly 7 Hours of Continuous Code Generation and Task Execution, Setting AI World Records, far surpassing its predecessor, Claude 3 Opus (less than 1 hour).
Anthropic claims that its AI models are not designed to eliminate jobs, but rather are a tool to automate everyday tasks. However, marktechpost media believes that the Claude 4 series will change the way AI is used, transforming it from a single-task aid to an "AI coworker" with stronger and broader capabilities.Can work almost a full work shift automatically.
Claude Sonnet 4: The universal choice for balancing performance and cost
Claude Sonnet 4 replaces its predecessor, Claude 3.5 Sonnet, with a more stable architecture that improves speed and quality without significantly increasing compute costs. The model is optimized for mid-scale deployments where there is a trade-off between cost and performance.
Although not as capable of reasoning as Opus 4, Sonnet 4 inherited many architectural upgrades, support for multi-file code navigation, intermediate tool usage and structured text processing, and better latency performance. It becomes the default model for Claude.ai's free users and is served via an API for lightweight development tools, user assistants, and analytics processes.
Technology Highlights and Deployment Methods
Both models have hybrid reasoning capabilities, offering Fast Mode for low-latency short dialog tasks, and Extended Thinking Mode for complex tasks requiring Extended Thinking Mode for complex tasks requiring deep reasoning and multi-round agent behavior.
This dual-mode strategy gives users the flexibility to allocate compute resources based on task complexity. In addition, Claude Opus 4 and Sonnet 4 are accessible through multiple cloud platforms including Anthropic's Claude API, Amazon Bedrock and Google Cloud Vertex AI, supporting a wide range of enterprise application scenarios from autonomous agents to code analysis.