Anthropic launches Claude Sonnet 5 as a cheaper way to run agents - BERITAJA

As shipping agentic capabilities becomes array stakes among instauration exemplary companies, Anthropic is releasing Claude Sonnet 5, a much powerful and agentic type of the lab’s midsize model. 

“It could make plans, usage devices for illustration browsers and terminals, and tally autonomously astatine a level that, conscionable a fewer months ago, required larger and much costly models,” Anthropic said successful a blog post. 

That framing mirrors what OpenAI and Google person said about their ain caller releases. OpenAI’s GPT-5.6 Sol was launched successful preview past week, and it is besides the firm’s about agentic exemplary yet, allowing users to divided activity crossed subagents for longer autonomous tasks. Google’s Gemini 3.5 Flash, which launched successful May, was sounded arsenic a displacement from a conversational chatbot to an agentic instrumentality that plans, builds, and iterates connected existent activity pinch minimal quality input.

Sonnet 5’s transportation is confirmation that agentic capacity is the caller baseline anticipation astatine each value tier. Now the differentiator isn’t going to beryllium who could do agentic activity best, but really cheaply they could do it and really reliably without quality oversight.  

Sonnet 5 promises capacity adjacent to that of Opus 4.8, but for overmuch little costs. Starting Tuesday, Claude Sonnet 5 will beryllium the default exemplary for free and Pro plans, and is disposable for each subscription.

At launch, Sonnet 5 is priced astatine $2 per cardinal input tokens and $10 per cardinal output tokens done August 31, aft which the value will jump to $3 per cardinal input tokens and $10 per cardinal output tokens. That makes Sonnet 5 cheaper than Opus 4.8, arsenic good arsenic OpenAI’s GPT-5.5 and Gemini 3.1 Pro. (It’s still much costly than Gemini 3.5 Flash.)

The caller exemplary besides demonstrates important improvements complete its predecessor Sonnet 4.6, released successful February, connected agentic capacity for illustration reasoning, instrumentality use, package coding, and knowledge work, according to Anthropic. 

For example, connected 1 benchmark, Sonnet 5 scores a 63.2% connected agentic coding, compared to Opus 4.8’s 69.2% and Sonnet 4.6’s 58.1%. On a knowledge activity benchmark, Sonnet 5 really somewhat outperforms Opus 4.8, which is known for winning connected solving the hardest problems for illustration making subtle judgement calls and heavy research. 

“Opus 4.8 is still the exemplary of prime for higher accuracy connected these tasks, but Sonnet 5 provides developers pinch lower-priced options that are of overmuch higher value than what was antecedently available,” Anthropic says. “Between Sonnet 5 and Opus 4.8, users could set the effort level to find the correct equilibrium of costs and performance.”

According to testers cited successful the blog post, Sonnet 5 besides excels astatine finishing analyzable tasks wherever erstwhile exemplary versions would person stopped short and “checks its ain output without explicitly being asked.”

“We handed Claude Sonnet 5 a two-part job—update Salesforce relationship tiers, nonstop a motorboat announcement to endeavor contacts—and it vanished extremity to end,” Daniel Shepard, a elder technologist astatine Zapier, said successful a statement. “That utilized to stall halfway. For day-to-day automation, it’s a no-brainer. ”

On safety, Sonnet 5 besides demonstrates a little complaint of “undesirable behaviors” for illustration practice pinch misuse and deception than its predecessor, making it safer to usage successful agentic contexts. It’s amended astatine refusing malicious requests and sidestepping hijack attempts successful punctual injection attacks. It besides hallucinates and engages successful sycophantic behaviour astatine a little complaint than Sonet 4.6.

That said, it’s not connected the aforesaid level arsenic Opus 4.8 and Claude Mythos Preview erstwhile it comes to misaligned behavior. “Evaluations besides show that it has a overmuch little expertise to execute vulnerable cybersecurity tasks than our existent Opus models,” sounds the blog post.

Lovable co-founder Fabian Hedin said successful a connection that Claude Sonnet 5 “refuses unsafe requests cleanly and consistently.”

“At Lovable, we’re putting powerful devices successful the hands of millions of builders,” Hedin said. “A exemplary that knows erstwhile to opportunity nary is conscionable arsenic important arsenic 1 that knows really to build.”