Anthropic launches Claude Sonnet 4.6: Near-Opus intelligence at lower cost, now default for Pro and Team users

Anthropic is moving fast. Just 12 days after introducing Claude Opus 4.6, its most powerful model, the company is back with another release — and this one may matter more for everyday users.

Claude Sonnet 4.6 is now live in Claude Code and has become the default model for Pro and Team subscribers. It’s cheaper than Opus 4.6, offers similar capabilities, and, in early testing, developers often chose it over Opus 4.5. That combination — price, performance, and preference — signals a shift in how Anthropic is positioning its models.

In a post on X, Anthropic said, “This is Claude Sonnet 4.6: our most capable Sonnet model yet. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also features a 1M token context window in beta.”

The company says Sonnet 4.6 performs strongly across coding, computer use, design tasks, knowledge work, and large-scale data processing. In plain terms, tasks that once required reaching for an Opus-class model can now run on Sonnet.

“Performance that would have previously required reaching for an Opus-class model — including on real-world, economically valuable office tasks — is now available with Sonnet 4.6,” Anthropic wrote in a blog post announcing the release.

That statement carries weight. Anthropic has built its brand on measured progress rather than splashy claims. When it suggests that Sonnet can handle economically valuable office work at near-Opus levels, it’s drawing a line between experimental AI and tools meant for daily use.

For free users and paid Pro users, Sonnet 4.6 now powers the Claude chatbot and the company’s Claude Cowork productivity tool. That decision matters. Default models shape adoption patterns. They influence what developers build on and what businesses test internally. When a company changes its default, it is signaling confidence.

Anthropic’s naming structure reflects its internal hierarchy. Opus sits at the top as the largest model. Sonnet occupies the middle tier. Haiku serves as the lightweight option. With Sonnet 4.6 narrowing the gap with Opus, the mid-tier is starting to look less like a compromise and more like a sweet spot.

The release arrives at a tense moment in the AI industry. Anthropic is competing head-to-head with OpenAI and Google, both of which are pushing out frequent model updates. Speed is no longer a differentiator; it is table stakes. Anthropic shipped Opus 4.6 on February 5. Less than two weeks later, Sonnet 4.6 is rolling out broadly.

There’s a broader market ripple here, too. Investors have grown uneasy about what advanced AI models mean for traditional software businesses. The iShares Expanded Tech-Software Sector ETF (IGV) has fallen more than 20% year to date, reflecting concerns that AI tools capable of writing code and automating office tasks could erode market share for established vendors, CNBC reported.

Sonnet 4.6 won’t calm those nerves. Anthropic says the model delivers “much-improved coding skills,” with more consistent performance and better adherence to instructions. Coding reliability is one of the benchmarks that separates novelty from production use. If a mid-tier model can handle that workload at scale, pricing pressure across the industry follows.

Anthropic itself is on solid financial footing. The company, founded in 2021 by former OpenAI researchers and executives, recently closed a $30 billion funding round at a $380 billion post-money valuation — more than double its September valuation. That places it among the most valuable private AI companies worldwide.

OpenAI is reportedly in talks to raise a round that could reach $100 billion. Google continues investing heavily in its Gemini models. The capital arms race is feeding a product arms race.

With Sonnet 4.6 becoming the default for Pro and Team users, Anthropic is making a clear bet: developers and businesses may not always need the biggest model. They need one that is close enough in intelligence, cheaper to run, and reliable on real tasks.

If early developer preference holds, Sonnet could become the workhorse of Anthropic’s lineup — the model people reach for daily rather than occasionally. And in this phase of AI, the default often wins.