Anthropic unveils Claude Opus 4.7 with advanced capabilities

Anthropic Expands Its Model Range with Claude Opus 4.7

Anthropic has recently introduced a significant update to its language models with the launch of Claude Opus 4.7. This new model stands out with the integration of the thinking_effort: xhigh option, which promises to greatly enhance performance in processing complex data.

Functional Innovations

Among the new features, two boolean options, thinking_display and thinking_adaptive, have been added. These functionalities aim to enrich the user experience by allowing summarized outputs available exclusively in JSON format or in JSON logs.

Additionally, Anthropic has increased the maximum number of max_tokens by default, now reaching the maximum allowed for each model. This change aims to optimize the efficiency and flexibility of the models during use.

Recent Changes and Events

Another notable change is the retirement of the outdated header structured-outputs-2025-11-13 for older models, signaling an update towards more modern practices.

Anthropic has also announced its participation in PyCon US 2026 in Long Beach, scheduled for April 17, 2026. This event will highlight new avenues in AI and security, reflecting the company's ongoing commitment to these areas.

Other Industry Announcements

At the same time, Meta's Muse Spark model has been unveiled, accompanied by new tools on meta.ai chat. These developments, announced on April 8, 2026, demonstrate the rapid evolution of the language model sector.

Support and Subscriptions

Simon Willison, the author of this article published on April 16, 2026, offers a monthly subscription service for $10 to receive a summary of the most important developments in language models. This initiative aims to keep subscribers informed of advancements in this constantly evolving field.