The AI industry moves at a pace that can feel exhausting, but Anthropic’s latest move is hard to ignore. Just twelve days after launching their powerhouse model, Claude Opus 4.6, the company has released Claude Sonnet 4.6. This new mid-tier model is designed to do something remarkable: deliver “Opus-level” intelligence at a fraction of the cost.
Claude Sonnet 4.6 brings flagship-level performance
In the world of AI, there has always been a trade-off. If you wanted a model capable of complex reasoning and flawless coding, you had to pay a premium for “Opus-class” models. If you wanted something cheaper, you had to settle for less “brain power.”
Claude Sonnet 4.6 is designed to effectively kill that trade-off. Anthropic has kept the pricing at $3 per million input tokens—the same as the previous version—despite the massive jump in capability. To put that in perspective, running a high-performance AI agent is now roughly five times cheaper than it was just a few weeks ago. For enterprises running millions of automated tasks a day, this is a big deal.
Better at “being human” (on a computer)
One of the most impressive leaps in this version is “computer use.” This is the ability of an AI to look at a screen, move a cursor, and type into apps just like a person would. When Anthropic first showed this off in late 2024, it was experimental and a bit clunky.
Sonnet 4.6 has reached a near-human level of proficiency, according to the company. In benchmarks designed to test how well AI can navigate web and desktop apps, it scored 72.5%, a sweet jump from the 61.4% of its predecessor. It is now so good at using a computer that it practically matches the much more expensive Opus 4.6 flagship. This is a game-changer for automating “legacy” tasks—those old databases and insurance portals that don’t have modern APIs.
The new king of coding?
For developers, Claude has already become a favorite tech tool, and Sonnet 4.6 doubles down on that reputation. In real-world testing, developers preferred its outputs 70% of the time over previous versions. It is reportedly less prone to “laziness” and follows complex, multi-step instructions with much higher consistency.
Beyond just writing code, the model showed it can “think long-term.” In a simulated business competition called “Vending-Bench,” the AI managed a business for a full simulated year. It developed a strategy of investing heavily in the first ten months to build capacity before pivoting to maximize profit. It ended the simulation with more than double the balance of previous models, proving it can plan for months, not just minutes.
Competitors like Google and OpenAI show unpredictable release dates and delays. Meanwhile, Anthropic is sticking to a steady, four-month update cycle. This consistency allows CTOs and developers to plan their roadmaps with the confidence that a better, more efficient model is always just around the corner. Anthropic can use this as a competitive advantage for the enterprise market.
Claude Sonnet 4.6 is now the default for everyone—including those on the free tier.
The post Anthropic Launches Claude Sonnet 4.6: Budget Price, Flagship-Level Performance appeared first on Android Headlines.
