Anthropic launched Claude Sonnet 5 globally Tuesday, introducing a new Sonnet-class artificial intelligence model with significantly enhanced coding and autonomous agentic performance.
The new model delivers capabilities previously associated with larger, more expensive models, making advanced multi-step task execution more accessible and efficient for professional applications.
Sonnet 5 is designed to autonomously handle complex, multi-step tasks, including planning and utilizing tools such as web browsers and terminals, Anthropic said. This development marks a substantial improvement in the model’s ability to act as an independent agent.
Anthropic stated that Sonnet 5 offers “near-Opus intelligence” for coding, agentic workflows and professional tasks, positioning it as a powerful tool despite being less capable than its flagship Opus and Mythos models.
The company highlighted Sonnet 5’s improved efficiency in token usage and introduced an introductory pricing structure of $2 per million input tokens and $10 per million output tokens, effective until Aug. 31.
In benchmark tests, Sonnet 5 demonstrated superior performance compared to its predecessor, Sonnet 4.6, as well as rival models like GPT-5.5 and Gemini 3.5 Flash. These benchmarks included BrowseComp, Terminal-Bench 2.1, SWE-bench Pro and FrontierCode.
The release underscores a competitive push in the generative AI sector, where companies like Anthropic and Cognition are continually advancing model capabilities to meet growing enterprise demand for more sophisticated and autonomous AI systems.
Anthropic emphasized that the model’s advancements in coding capabilities are particularly notable, offering developers a more robust tool for complex programming challenges.
The global release of Claude Sonnet 5 is expected to affect various industries by providing a more powerful and cost-effective solution for automating intricate digital processes and supporting advanced professional work.
Source: Search Engine Journal
