Anthropic has unveiled Claude 3.7 Sonnet, its most advanced AI model to date and the first hybrid reasoning model on the market.
Claude 3.7 Sonnet offers a unified model that can operate in both standard and extended thinking modes. This approach allows organisations to deploy a single AI system across varying business needs rather than managing multiple specialized models.
"Claude 3.7 Sonnet is both an ordinary LLM and a reasoning model in one: you can pick when you want the model to answer normally and when you want it to think longer before answering," Anthropic states in its announcement.
The model is immediately available across all Claude plans—Free, Pro, Team, and Enterprise—as well as through Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI. Extended thinking capabilities are available on all tiers except the free version. Pricing remains consistent with previous models at $3 per million input tokens and $15 per million output tokens, including thinking tokens.
A key innovation for business implementations is the API's granular control over reasoning resources. Users can specify exactly how many tokens (up to 128K) Claude should use for thinking, allowing precise balancing of cost, performance, and response time for enterprise deployments. This level of control enables businesses to establish consistent performance benchmarks and predictable operational costs across different AI use cases.
The most significant enterprise impact appears to be in software development, where Claude 3.7 Sonnet has achieved state-of-the-art performance on multiple industry benchmarks. The model demonstrated superior results on SWE-bench Verified, which tests AI systems' ability to resolve real-world software issues, and TAU-bench, which evaluates complex task completion with user and tool interactions.
Alongside the new model, Anthropic introduced Claude Code, an agentic command line tool for developers, available in limited research preview. This tool enables developers to delegate substantial engineering tasks directly from their terminal, with capabilities including searching and reading code, editing files, writing and running tests, and committing code to GitHub.
In early implementation testing, "Claude Code completed tasks in a single pass that would normally take 45+ minutes of manual work, reducing development time and overhead." Anthropic plans ongoing improvements, including enhanced tool call reliability, support for long-running commands, improved rendering, and expanded capabilities.
Enterprise developers can also access improved GitHub integration, now available across all Claude plans. This integration enables direct connection to code repositories and allows for more efficient bug fixing, feature development, and documentation building across personal, work, and open source projects.
Claude 3.7 Sonnet's hybrid approach addresses a fundamental challenge for enterprise AI implementation: balancing computational efficiency with analytical depth. Organisations can now deploy a single model that handles both quick operational responses and complex analytical problems, reducing both technical overhead and training requirements.
Early testing with partners demonstrates quantifiable productivity improvements for software development teams, with particular strength in complex codebase navigation, full-stack updates, and automated test development—often reducing development time for certain tasks from hours to minutes.
The model also shows a 45% reduction in unnecessary refusals compared to previous versions, addressing a common enterprise complaint about overly cautious AI systems that limit practical business utility.