San Francisco, California — Anthropic has unveiled its latest iteration of Large Language Model technology, Claude Sonnet 4.6. Released on Tuesday, this newest model follows closely behind the launch of Claude Opus 4.6, which debuted on February 5. According to the company, this latest Sonnet model is its most advanced yet, with a beta version that boasts a 1 million token context window.
Anthropic emphasizes the model’s safety features, stating that Claude Sonnet 4.6 exhibits reduced instances of generating false information and displays less tendency toward flattery. The enhanced coding capabilities built into this model are particularly appealing for developers who rely on AI tools for software creation.
Access to Claude Sonnet 4.6 is straightforward. The model is now the default option for both free and Pro users through the company’s platforms, including claude.ai and Claude Cowork. Additionally, it is available via the API on major cloud service providers.
Free users can interact with the model, albeit with limited usage constrained by demand, with limits resetting every five hours. For those requiring greater access, the Claude Pro subscription remains at $20 per month or $17 when paid annually. For API users, pricing begins at $3 per million input tokens and $15 per million output tokens.
Benchmark tests reveal that Claude Sonnet 4.6 outperforms its predecessors and rivals in specific tasks, particularly in financial analysis and office-oriented assignments. It has surpassed notable competitors such as Google’s Gemini 3 Pro and OpenAI’s GPT 5.2. Surprisingly, throughout these assessments, Claude Sonnet 4.6 even outshone Anthropic’s own Opus 4.6 model in several categories.
Early users of Claude Sonnet 4.6 have reported a preference for this new model over both Claude Sonnet 4.5 and Claude Opus 4.5. The system card for Sonnet 4.6 indicates improvements in several key metrics, including scores on Humanity’s Last Exam, though Opus 4.6 achieved a higher overall score.
In one significant testing case, an AI-driven insurance company noted that Sonnet 4.6 led the field against all previous Claude models on its intricate insurance benchmark, underscoring its versatility. While Claude Opus models typically excel in complex reasoning tasks, the competitive pricing and performance of Claude Sonnet 4.6 provide an attractive alternative for users seeking effective AI solutions without breaking the bank.
As the landscape of artificial intelligence continues to evolve, Anthropic’s latest release represents a significant step forward in accessibility and performance, catering to an increasing demand for innovative AI applications across various industries.