Claude Sonnet 5.5 is faster and up to 30 per cent cheaper per task

Anthropic has released Claude Sonnet 5.5, the second model in its Claude 5.5 family, and is selling it on speed and running cost rather than raw power. The company says it is a clear upgrade over Sonnet 5, produces output more than 30% faster and costs up to 30% less for most work.[1]

The list price has not moved. The saving, Anthropic says, comes from the model needing fewer tokens to finish the same job. All performance and cost figures in the announcement are the company's own or come from testers it chose to quote, and none has been independently audited.[1]

Sonnet 5.5 sits below Claude Opus 5.5, which Anthropic released earlier this month for complex work that needs careful judgment. Anthropic says Sonnet is strongest at well-scoped everyday tasks such as fixing bugs and producing documents, slides and spreadsheets. A third model, Claude Haiku 5.5, aimed at high-volume and cost-sensitive work, is due in the coming weeks.[1]

Anthropic is open about the gap with its top model. On several benchmarks, Sonnet 5.5 at its highest effort setting scores close to Opus 5.5, but the company says Opus remains clearly stronger at open-ended work that needs sustained judgment. Benchmark scores, it adds, capture only one facet of what a model can do.[1]

On CursorBench 4.0, built from real Cursor coding sessions, Anthropic reports 55.5% for Sonnet 5.5, against 34.1% for Sonnet 5 and 57.8% for Opus 5.5. On the OSWorld 2.1 computer-use test the scores are 80.1%, 57.0% and 81.8%. On GDPval-AA, covering tasks from 44 occupations, Sonnet 5.5 is about 400 points above Sonnet 5.[1]

The cost claim rests on effort settings. Anthropic says Sonnet 5.5 at low or medium effort beats Sonnet 5's best score on several benchmarks for about a tenth of the cost per task. Medium is the default in the Claude apps and Claude Code. Balyasny Asset Management, one quoted tester, reported about 121,000 tokens per finance answer against 497,000 for Sonnet 5.[1]

Because Sonnet 5.5's cybersecurity skills are comparable to Opus 5's, Anthropic says it is the first Sonnet model to launch with cyber safeguards like those on its most capable models. Routine bug finding and fixing is unaffected, but higher-risk security tasks will visibly fall back to Sonnet 5. Its biology safeguards are unchanged.[1]

It is also the first Sonnet with classifiers meant to block distillation, where attackers use thousands of fake accounts to copy a model's abilities. On Anthropic's automated audit of about 1,850 scenarios, Sonnet 5.5 matched or improved on Sonnet 5 on most alignment measures. The company says no set of tests catches every failure.[1]

Malaysia is on Anthropic's list of supported countries, and Sonnet 5.5 is available on Anthropic's own platform as well as Amazon Web Services, Google Cloud and Microsoft Azure. Billing is in US dollars. At Bank Negara Malaysia's 29 September rates of RM4.065 to RM4.09, input costs roughly RM8.20 per million tokens and output roughly RM41.[2][1][3]

For SMEs and agencies already paying for Sonnet 5, the practical change is that the same budget may stretch further, if the token savings hold on their own work. Developers who run Sonnet with thinking switched off must first move to a new setting called between_tools, according to Anthropic's migration note.[1]