Claude Sonnet 5 Now Handles Multi-Step Tasks Without Paid Upgrades

The real story in Anthropic’s latest Claude update is not the speed bump or the sharper coding answers. It is that Sonnet 5 now handles autonomous, multi-step tasks at a level that previously required larger and more expensive models, according to Anthropic’s own announcement. That is a meaningful shift in where the capability ceiling sits for a mid-tier model.

Anthropic describes Sonnet 5 as its “most agentic Sonnet yet,” meaning the model can make plans, use tools like browsers and terminals, and run through tasks without constant hand-holding. Until recently, that kind of autonomous behavior was the domain of the company’s heavier, pricier offerings. Folding it into what is now the default model changes the value equation for anyone who was paying up for that capability.

The update also claims better memory within a conversation, so users working through long projects should need to re-explain context less often. Combined with the faster response times, the practical effect is a chatbot that feels less like issuing commands to a machine and more like working with something that tracks where you left off.

Access and availability

Rollout is already live. Anyone logging into the Claude web interface or opening the iOS app gets Sonnet 5 as the default without changing any settings. Free users have immediate access, though they will hit daily message limits faster than subscribers on paid plans.

By making Sonnet 5 the automatic default rather than an opt-in, Anthropic is doing something straightforward: retiring the older versions from the front of the experience rather than letting them linger as a fallback. That keeps the product simple, but it also means users who preferred an older model’s behavior have no easy path back.