Anthropic launches Claude 3.7 Sonnet: The world’s first hybrid reasoning AI that blends speed with deep thinking

After weeks of speculation, AI startup Anthropic has introduced Claude 3.7 Sonnet, an AI model that combines quick responses with controlled, extended reasoning. Alongside it, the company has rolled out Claude Code, a command-line coding tool designed to simplify development workflows. The launch comes as Anthropic finalizes $3.5 billion in fresh funding, signaling its push for more advanced AI development.

“Introducing Claude 3.7 Sonnet: our most intelligent model to date. It’s a hybrid reasoning model, producing near-instant responses or extended, step-by-step thinking,” Anthropic said in a post on X.

Introducing Claude 3.7 Sonnet: our most intelligent model to date. It’s a hybrid reasoning model, producing near-instant responses or extended, step-by-step thinking.

One model, two ways to think.

We’re also releasing an agentic coding tool: Claude Code. pic.twitter.com/jt7qQmFWuC

— Anthropic (@AnthropicAI) February 24, 2025

A New Approach to AI Reasoning

Claude 3.7 Sonnet gives users the option to switch between standard and extended thinking modes, allowing for more detailed reasoning when needed. API users can fine-tune how long the model processes information, with support for up to 128K tokens, making it easier to balance speed, cost, and output quality.

Claude 3.7 Sonnet takes a major leap forward from its predecessor, introducing extended thinking mode, which sharpens its performance in math, physics, coding, and complex problem-solving.

API users now have fine-grained control over how long the model processes information, allowing them to balance speed, accuracy, and cost based on task requirements.

The model outperforms competitors like o1, o3-mini, and DeepSeek R1 in real-world coding benchmarks and AI-driven tool use. It’s available through the Claude app, Anthropic’s API, Amazon Bedrock, and Google’s Vertex AI—all at the same price as its predecessor: $3 per million input tokens and $15 per million output tokens.

With an updated knowledge base as of October 2024, Claude 3.7 Sonnet is well-suited for coding, finance, and legal analysis, giving businesses a more capable AI for tackling complex tasks.

Claude Code: A Hands-Free AI Assistant for Developers

Anthropic also introduced Claude Code, a command-line assistant currently in limited research preview. The tool can edit files, read code, run tests, and manage GitHub commits—all without requiring manual intervention. The goal is to reduce tedious development tasks and let engineers focus on higher-level problem-solving.

Agentic Coding

Claude Code represents a shift toward agentic coding, where AI moves beyond simply suggesting code and starts executing tasks. The tool can:

• Read and understand existing code to make precise modifications
• Edit files without requiring manual intervention
• Run tests, detect errors, and suggest fixes
• Handle GitHub commits and track changes automatically

By automating these repetitive processes, Claude Code reduces the manual workload for developers, allowing them to focus on more complex problem-solving.

Currently, in limited research preview, this tool signals Anthropic’s deeper push into AI-driven development, offering a glimpse into a future where AI can act as an independent coding assistant.

The Competitive Landscape

AI-driven reasoning is becoming the next major battleground, and Anthropic is making its move with Claude 3.7 Sonnet. The company had been relatively quiet in recent months, but this release puts it back in the spotlight. With DeepSeek pushing open-source models and OpenAI working on its own hybrid AI, the race is heating up.

Claude 3.7 Sonnet signals a shift for Anthropic—a more strategic, high-performance AI designed to give users greater control over reasoning and problem-solving. The next generation of AI models is here, and the competition is just getting started.

Here’s a YouTube video from Anthropic covering the launch of Claude 3.7 Sonnet and Claude Code.