Artificial intelligence giant Anthropic has officially launched its new large language model, Claude Sonnet 5. This model is positioned as a mid-to-high-end option with excellent cost-effectiveness, and its core advantage lies in significantly narrowing the performance gap with its top-tier flagship Opus series.

As the "most agent-like" Sonnet model to date, it not only performs well in reasoning, programming, and knowledge-based tasks, but can also autonomously plan complex tasks. Users do not need to provide additional instructions; the model can self-check its output results and flexibly use external tools such as browsers and terminals.

Low-cost strategy to reach more users

In terms of security, the new model also performs exceptionally well, with a significant reduction in hallucination rates and agreeable responses, and it is more intelligent in rejecting malicious requests. Currently, Sonnet 5 is officially available for all plans and has directly become the default large model for both the free plan and the Pro paid plan.

To quickly capture the market, Anthropic has shown great诚意 in pricing, with costs per million input and output tokens far lower than those of the flagship Opus model. The official has adopted this downgrading pricing strategy to enable a broader range of developers and general users to enjoy top-tier AI automation experiences at a very low threshold.