Claude Sonnet 4.6 delivers frontier-level AI for free and cheap-seat users

Anthropic / Elyse Betters Picaro / ZDNET

Follow ZDNET: Add us as a preferred source on Google.

ZDNET's key takeaways

• Claude Sonnet 4.6 is a major upgrade over 4.5.
• 1M-token context window (in beta) enables longer, richer sessions.
• It's now the default for free and Pro users, with pricing unchanged.

Just four months after Anthropic released Claude Sonnet 4.5, and a week after the company debuted Opus 4.6, the AI giant is back with Claude Sonnet 4.6, a huge upgrade over the previous version.

Also: Anthropic says its new Claude Opus 4.6 can nail your work deliverables on the first try

This new Sonnet 4.6 model, available now, shows improved coding performance, better computer use skills, upgraded long-context reasoning, better agent planning, and improvements to knowledge work and design.

As with Opus 4.6, Sonnet 4.6 now includes a 1 million-token context window (in beta). This allows for much longer and more complex work sessions without requiring a session reset or compaction.

Sonnet 4.6 is now the default model for free and Pro tier users across the various Claude interfaces. Pricing for those plans (as well as for Sonnet API use) has not increased.

Closing the gap with Opus

Anthropic provides two branded AI models at different price points, Sonnet and Opus. Opus has always been the Cadillac of AI models, available at higher tiers and increased per-token API call pricing. Sonnet has been more of an entry-level model, still quite capable, but with substantially lower resource usage, enabling Anthropic to deploy it to free users and keep its token price down.

According to the company's blog post announcing the release of Sonnet 4.6, "It approaches Opus-level intelligence at a price point that makes it more practical for far more tasks."

Also: I used Claude Code to vibe code a Mac app in 8 hours, but it was more work than magic

According to the company's testing, performance that previously would have only been seen in an Opus-class model is now available for users of Sonnet 4.6. This new model also shows major improvements in AI-based desktop computer interaction.

There are some practical limits, however. The company says, "The model certainly still lags behind the most skilled humans at using computers. But the rate of progress is remarkable nonetheless. It means that computer use is much more useful for a range of work tasks, and that substantially more capable models are within reach."

Developers prefer it, even over Opus 4.5

In early user testing, Anthropic found that developers preferred Sonnet 4.6 over Sonnet 4.5 about 70% of the time. The company says, "Users reported that it more effectively read the context before modifying code and consolidated shared logic rather than duplicating it. This made it less frustrating to use over long sessions than earlier models."

I am curious about that remaining 30% though. You'd think with an apples-to-apples upgrade like Sonnet 4.5 to 4.6 that nearly all users would prefer the newer model. I've asked Anthropic why the remaining 30% presumably didn't favor the new release. Stay tuned. If I learn anything, I'll share it here.

Also: Claude Code made an astonishing $1B in 6 months - and my own AI-coded iPhone app shows why

When comparing Sonnet 4.6 to Opus 4.5 (the older frontier model released in November), developers preferred Sonnet 4.6 roughly 60% of the time.

The company reported that early users, "Rated Sonnet 4.6 as significantly less prone to overengineering and laziness, and meaningfully better at instruction following. [Early users] reported fewer false claims of success, fewer hallucinations, and more consistent follow-through on multi-step tasks."

Given that the current general-availability version of Opus is 4.6, this result isn't a harbinger of a mass migration off of the Opus model by higher-tier users. But what it does say is that the "cheap seats" model has improved enough to be up to tasks previously reserved for higher-performing models.

Speed, practicality, and everyday workflows

Let's not underestimate the benefits of the higher performance, yet lower resource usage, that Sonnet 4.6 shows. When using the free and Pro tiers, Anthropic will throttle usage based on token use and resource usage. Sonnet 4.6's improvements are akin to a car getting more miles per gallon when using a new gasoline, especially if the "pickup 'n go" is still as good or better.

Also: 10 things I wish I knew before trusting Claude Code to build my iPhone app

The four-times-larger 1-million-token window also provides a practical benefit. It can hold entire codebases, lengthy contracts, or dozens of research papers. Anthropic says, "More importantly, Sonnet 4.6 reasons effectively across all that context. This can make it much better at long-horizon planning."

Where Opus 4.6 still leads

Don't give up on Opus, however. Opus 4.6 is still Anthropic's frontier model champion.

Also: I stopped using ChatGPT for everything: These AI models beat it at research, coding, and more

The company says, "We find that Opus 4.6 remains the strongest option for tasks that demand the deepest reasoning, such as codebase refactoring, coordinating multiple agents in a workflow, and problems where getting it just right is paramount."

A practical default for daily work

Anthropic is positioning Sonnet 4.6 as a practical daily driver. In many cases, it's considerably faster than Opus 4.6.

In that way, there are clear competitive parallels between OpenAI's GPT-5.3-Codex-Spark and its GPT-5.3-Codex, with Spark the faster and less accurate version and the full Codex the frontier model leading development. One big difference is that while Anthropic says Sonnet 4.6 is faster, it's not making anything like the 15x performance claim that OpenAI made of its Spark model.

Also: Which AI tools are actually worth paying for? I'm keeping these subscriptions in 2026 - here's why

(Disclosure: Ziff Davis, ZDNET's parent company, filed an April 2025 lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

For most coding and knowledge work, Sonnet 4.6 offers strong performance, particularly for those on the lower pricing tiers. It also offers a solid price/performance profile for users working with API calls who want to get as much bang for the buck as possible. Meanwhile, Opus 4.6 remains a viable escalation path for more complex problems needing deeper reasoning.

What about you? Have you tried Claude Sonnet 4.6 yet? If so, how does it compare to Opus in your real-world workflows? Does the 1-million-token context window change how you approach coding, research, or long planning sessions?

Are you comfortable relying on the "cheap seats" model for serious work, or do you still escalate to Opus for high-stakes tasks? And if you're on the free or Pro tier, do these improvements make you more likely to stick with Sonnet as your daily driver? Let us know in the comments below.

You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.