Anthropic’s First Reasoning AI Model is Now Available

Anthropic today announced the immediate availability of Claude 3.7 Sonnet, its first reasoning AI model, to all Claude plans, including free. It’s particularly well-suited for software coding and front-end web development, the company says, and so it has also released a command line agentic coding tool called Claude Code as a limited research preview.

“We’ve developed Claude 3.7 Sonnet with a different philosophy from other reasoning models on the market,” Anthropic notes in the announcement post. “Just as humans use a single brain for both quick responses and deep reflection, we believe reasoning should be an integrated capability of frontier models rather than a separate model entirely. This unified approach also creates a more seamless experience for users.”

According to Anthropic, Claude 3.7 Sonnet is both an “ordinary” large language model (LLM) and a reasoning model, and users can switch between the two modes on the fly. In its standard mode, the model is like an upgraded version of Claude 3.5 Sonnet, but in extended thinking mode, it will “self-reflect” before answering, improving its performance with math, physics, instruction-following, coding, and many other tasks. Like other reasoning models, Claude 3.7 Sonnet makes its “thought process” visible to the user in that second, extended thinking mode.

Developers who access Claude 3.7 Sonnet through its API can target either mode as well, and Anthropic provides a new budget feature so they can compromise between speed and performance as desired. The firm also notes that it has started shifting Claude away from math and computer science competition problems with this release and to real-world tasks that will better benefit typical businesses. That said, Anthropic claims that Claude 3.7 Sonnet is “once again best-in-class for real-world coding tasks, with significant improvements in areas ranging from handling complex codebases to advanced tool use.”

Developers interested in working with Claude 3.7 Sonnet can do so now via Claude.ai and GitHub, and across Anthropic API, Amazon Bedrock, and Google Cloud Vertex AI. You can learn more on the Anthropic Docs website.

Claude Code, meanwhile, is Anthropic’s first agentic coding tool and it’s available only in a limited research preview.

“Claude Code is an active collaborator that can search and read code, edit files, write and run tests, commit and push code to GitHub, and use command line tools—keeping you in the loop at every step,” the company explains. “Claude Code is an early product but has already become indispensable for our team, especially for test-driven development, debugging complex issues, and large-scale refactoring. In early testing, Claude Code completed tasks in a single pass that would normally take 45+ minutes.”

Instructions for signing up for Claude Code are available on the Anthropic Docs website.