Related Articles
Claude Sonnet 4, who was released in Anthropic, has been paraphrased. The most powerful Sonnet model to date, according to 6, is the major improvements in coding, long context reasoning (as well as other large-scale changes), agent planning, computer use and design.
A 1 million token context window is introduced in beta, allowing users to process entire codebases, long contracts or large research archives within one prompt.
Sonnent 4, however, is not charged with any changes in pricing from Sonnet 4. Sonnet 4, Sonned 4 and 5 at $3 input and $15 output per million tokens,. Free and Pro tiers now default model is 6 on the default models of . Sonnet tier Anthropic says performance that had been required for an Opus class model is now available in the performer’s line-up, which was previously performed by Aristopic.
Sonnet 4, which was popular with early users, drew on Sonnel 4. Sonnet 4 6 over Sonnetion 4 . 5 about 70% of the time and even preferred it over Claude Opus 4 in favour of . A fifth in 59% of tests, citing better instruction after the test, less hallucinations and reduced overengineering, was 5 (in 59).
A similar model also shows material advances on the computer’s use of . Sunet 4 (on the OSWorld benchmark) – which tests real software interaction without APIs, Sonnet 4, is on the test of this standard. Six approaches human-level performance on tasks like navigating spreadsheets and performing multi-step web workflows. The reissue includes increased resistance to prompt injection attacks, Anthropic said.
launch follows Anthropic’s early February unveiling of its most advanced model, Claude Opus 4. 7, placing Sonnet 4. A cheaper alternative to 6 is the ability gap that narrows down on capabilities, reducing the capability gap and being less expensive.
Thanks for reading Anthropic launches Claude Sonnet 46 with 1M token context window
