As the race to build increasingly autonomous artificial intelligence intensifies, Anthropic has introduced Claude Sonnet 5, a new mid-sized language model designed to make AI agents significantly more capable and affordable.
The launch reflects a broader shift across the AI industry, where competition is no longer centered solely on chatbot performance but on the ability of AI systems to complete complex, multi-step tasks with minimal human intervention. Anthropic says Claude Sonnet 5 delivers the kind of autonomous performance that, until recently, required much larger and considerably more expensive models.
Claude Sonnet 5 replaces the company’s previous Sonnet model and is positioned as the default option for most developers and enterprise customers building AI-powered applications. According to Anthropic, the model has been engineered specifically for “agentic” workflows—applications in which AI can independently plan tasks, use external tools, browse websites, interact with software, write and execute code, and adapt its actions based on changing information.
The company believes the latest model represents a major leap in practical AI deployment. Rather than simply responding to prompts, Claude Sonnet 5 is designed to function as an intelligent assistant capable of completing lengthy workflows that previously required continuous user supervision. Anthropic says the model can effectively utilize browsers, command-line terminals, APIs, and other digital tools to accomplish real-world objectives.
One of the biggest selling points is cost. During its introductory period, Claude Sonnet 5 is priced at $2 per million input tokens and $10 per million output tokens—significantly lower than existing AI models. After August 31, the pricing will increase to $3 per million input tokens and $15 per million output tokens, but Anthropic says the model will remain one of the most cost-effective options for enterprise AI agents.
The aggressive pricing strategy highlights the growing competition among AI developers. Over the past year, companies including OpenAI, Google, xAI, and Anthropic have increasingly focused on reducing the cost of deploying advanced AI systems while simultaneously improving their reasoning and automation capabilities. Lower inference costs are becoming a crucial factor for businesses building AI products that may execute thousands—or even millions—of automated tasks every day.
Anthropic says Claude Sonnet 5 approaches the performance of its much larger flagship models on many real-world tasks while consuming fewer computational resources. Early demonstrations showcased the model independently completing workflows such as updating customer information in enterprise software, managing browser-based tasks, and sending communications without requiring human intervention between steps.
Developers stand to benefit from the model’s expanded capabilities. Anthropic says Claude Sonnet 5 can execute more sophisticated coding tasks, better manage long-running projects, and maintain coherence during complex reasoning exercises. The company has also strengthened its computer-use abilities, allowing the model to interact more naturally with desktop environments and web applications.
Claude Sonnet 5 is now available through Anthropic’s API and has become the default model for users across Claude’s Free and Pro subscription tiers, with access also extending to Team, Max, and Enterprise customers. The rollout signals Anthropic’s confidence that agentic AI is moving beyond experimentation and into mainstream business adoption.
