Introducing Claude Opus 4.7
Summary (EN)
Anthropic announced general availability of Claude Opus 4.7 on April 16, presenting it as an upgraded version of Opus 4.6 with stronger performance on advanced software engineering and long-running, multi-step tasks. Anthropic said the model improves on complex coding work that previously required close human supervision, with gains in instruction following, consistency over extended task horizons, and self-verification before returning answers. The company also highlighted better vision performance, including higher-resolution image understanding, and said the model produces stronger interfaces, documents, and presentation-style outputs. In parallel with capability improvements, Anthropic framed Opus 4.7 as part of its staged rollout strategy following Project Glasswing and the restricted unveiling of Claude Mythos Preview. The company said Opus 4.7 is the first less-capable model on which it is deploying new safeguards intended to detect and block prohibited or high-risk cybersecurity use cases, while inviting legitimate security professionals into a Cyber Verification Program for approved defensive work. Anthropic kept pricing unchanged from Opus 4.6 and said Opus 4.7 is available across Claude products, the API, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry. The release is notable both as a major frontier-model update for developer and enterprise workflows and as a live test of Anthropic’s attempt to combine higher autonomy with tighter cyber-risk controls before any broader release of Mythos-class systems.
Summary (ZH)
Anthropic 于 4 月 16 日正式宣布 Claude Opus 4.7 全面可用，将其定位为 Opus 4.6 的升级版本，重点提升高级软件工程、长时程任务执行和多步骤复杂工作流能力。Anthropic 表示，新模型在复杂编程任务上的表现明显增强，能够更稳定地遵循指令，在长任务链中保持一致性，并在输出前主动进行结果校验。除了编码场景，官方还强调其视觉能力提升，支持更高分辨率图像理解，并在界面生成、文档撰写和演示材料等专业任务中表现出更好的质量与风格控制。此次发布同时被纳入 Anthropic 在 Project Glasswing 与 Claude Mythos Preview 之后的分阶段部署策略中。Anthropic 表示，Opus 4.7 是首个搭载新型网络安全防护机制的较低风险模型，这些机制会自动识别并阻断涉嫌违规或高风险的网络安全请求；对于合规的防御性用途，安全研究人员则可通过 Cyber Verification Program 申请使用。价格方面维持与 Opus 4.6 相同，产品已在 Claude 全系产品、Anthropic API、Amazon Bedrock、Google Cloud Vertex AI 和 Microsoft Foundry 上线。此次更新的重要性在于，它既是面向开发者和企业工作流的重要前沿模型迭代，也是 Anthropic 在更高自治能力与更严格网络安全防护之间进行真实部署验证的一次关键产品动作。