New Coding Model from Anthropic

<p>Anthropic has released Claude Opus 4.5 - a model that outperforms all existing AI models for coding, agents, and complex workflows in tests. This is a new peak in how AI understands code, eliminates bugs, and manages multi-system tasks without manual oversight.</p>
<p>Opus 4.5 confidently addresses real engineering problems: on the SWE-bench Verified, it shows results higher than not only other models but also strong engineers in Anthropic's tests. Improvements include reasoning, handling large contexts, and resilience to prompt injection. Essentially, this makes the model a more reliable tool for development.</p>
<p>Anthropic has also updated the Developer Platform: advanced tool use, effort parameter, and new context management allow for building long-lived agent chains that operate more efficiently and cost-effectively. Claude Code has been upgraded and is now available in the desktop application.</p>
<p>Source: <a href="https://www.anthropic.com/news/claude-opus-4-5">https://www.anthropic.com/news/claude-opus-4-5</a></p>;
<p>#ai #agi #anthropic #coding</p>