Anthropic Launches Claude 3.7 Sonnet, Its First Hybrid Reasoning Model

Anthropic announced on February 24 the launch of Claude 3.7 Sonnet, which it calls its most intelligent model to date and, the company says, the first hybrid reasoning model on the market. The news isn't just that it reasons better, but how it does it: a single model that can answer almost instantly or pause to think step by step, showing that process to the user. Alongside it comes Claude Code, a command-line tool for delegating engineering tasks directly from the terminal.

One brain for fast answers and slow thinking

Anthropic's approach starts from an idea deliberately different from other labs. Rather than keeping a separate model for reasoning and another for quick replies, the company argues that reasoning should be an integrated capability of frontier models rather than a separate model entirely. Just as humans use a single brain for both quick responses and deep reflection, Anthropic reasons, a frontier model should do the same.

In practice, Claude 3.7 Sonnet functions as two things at once. In its standard mode, it's an upgraded version of the previous Claude 3.5 Sonnet: it answers directly. In extended thinking mode, the model reflects before answering, which Anthropic says improves its performance on math, physics, instruction-following, coding, and other tasks. Users decide when they want an immediate answer and when they'd rather the model take its time.

That choice isn't just an on/off switch. Through the API, developers can set a thinking budget: telling Claude to think for no more than a set number of tokens, up to its output limit of 128,000 tokens. It's a fine-grained control that lets you trade off speed and cost against answer quality. The longer you let it think, the more expensive and slower it gets—but potentially better.

Fewer olympiads, more real work

There's a notable nuance in how Anthropic trained this model. The company says it optimized somewhat less for math and computer science competition problems—the kind of olympiad-style challenges labs love to tout in announcements—and shifted focus instead toward real-world tasks that better reflect how businesses actually use these models.

It's a distinction worth highlighting. Much of the race between labs has played out on synthetic leaderboards that impress but say little about day-to-day use. Prioritizing real-world work amounts to acknowledging that the value of these systems is measured in concrete tasks, not contest scores.

Coding: where the leap is most visible

The area where Claude 3.7 Sonnet shows its clearest gains is coding and front-end web development. Anthropic says it achieves state-of-the-art performance on SWE-bench Verified, a benchmark that evaluates models' ability to solve real-world software issues, as well as on TAU-bench, a framework that tests AI agents on complex real-world tasks involving user and tool interactions.

Praise from companies that tested it ahead of launch points in the same direction. Cursor noted Claude 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. Cognition found it far better than any other model at planning code changes and handling full-stack updates. Vercel highlighted Claude's exceptional precision for complex agent workflows, while Replit has successfully deployed Claude to build sophisticated web apps and dashboards from scratch, where other models stall. In Canva's evaluations, Claude consistently produced production-ready code with superior design taste and drastically reduced errors.

That pile-up of customer testimonials is marketing, worth keeping in mind. But it lines up with a stat Anthropic keeps repeating: since mid-2024, Sonnet has been the preferred model among developers worldwide. Coding has become the use case where these models demonstrate the most measurable utility.

Claude Code: an agent that lives in the terminal

The second announcement is Claude Code, Anthropic's first agentic coding tool, arriving as a limited research preview. The key word is agentic: this isn't an assistant that suggests code snippets, but an active collaborator that acts.

According to the company, Claude Code can search and read code, edit files, write and run tests, commit and push code to GitHub, and use command line tools—keeping the developer in the loop at every step. Anthropic says it has already become indispensable for its own team, especially for test-driven development, debugging complex issues, and large-scale refactoring. In early testing, the company says, it completed tasks in a single pass that would normally take 45+ minutes of manual work.

The program's stated goal is to learn. Anthropic wants to understand how developers use Claude for coding in order to inform future model improvements. In the coming weeks, it plans to improve tool call reliability, add support for long-running commands, and enhance in-app rendering.

Anthropic has also improved the coding experience on Claude.ai: its GitHub integration is now available on all plans, letting developers connect their repositories directly to the model to fix bugs, build features, and write documentation.

Price and availability, no surprises

One detail that doesn't usually accompany capability leaps: the price stays the same. In both standard and extended thinking modes, Claude 3.7 Sonnet costs the same as its predecessors—$3 per million input tokens and $15 per million output tokens, a figure that includes thinking tokens. In other words, extended reasoning consumes output tokens and is billed as such.

The model is available on all Claude plans—Free, Pro, Team, and Enterprise—as well as on Anthropic's developer platform, on Amazon Bedrock, and on Google Cloud's Vertex AI. Extended thinking mode is available on all surfaces except the free tier.

Safety: fewer refusals, new risks

Anthropic says it put the model through extensive testing with external experts. One concrete change: Claude 3.7 Sonnet draws a more nuanced distinction between harmful and harmless requests, cutting unnecessary refusals by 45% compared with the previous version. That's a practical fix for a known problem—models that reject legitimate tasks out of excessive caution.

The system card accompanying the launch also addresses the emerging risks of AI computer use, particularly prompt injection attacks—when a third party embeds hidden malicious instructions for the model to execute—and explains how Claude is trained to resist them. It also examines an interesting angle of reasoning models: the ability to see how they make decisions and to assess whether that reasoning is actually trustworthy.

What it means

The combination of a hybrid model with thinking-budget controls and an agent that operates in the terminal points to a different way of working for developers. The promise is to delegate substantial engineering tasks, not just autocomplete lines of code. That Anthropic is launching it as a limited preview, acknowledging it's an early product with reliability still to be polished, fits the caution of a company that knows letting an agent commit and push code to GitHub is as powerful as it is delicate.

The real insight here is the single-model philosophy: instead of forcing users to choose between a fast model and one that reasons, Anthropic leaves that decision open question by question, and even lets you dial it in by tokens. It remains to be seen whether extended thinking earns its cost on everyday tasks, or whether, as usually happens, most real work gets handled just fine in fast mode.