Anthropic Unveils Claude Opus 4.7 With Advanced Agentic Capabilities

Daily Technology

17/04/2026

Anthropic has announced the release of its latest artificial intelligence model, Claude Opus 4.7. Positioned as a significant upgrade over its predecessor, this new iteration focuses on enhancing performance in complex professional tasks, particularly in coding and engineering.

Performance Benchmarks and Key Improvements

The primary advancements in Claude Opus 4.7 are evident in its agentic coding capabilities. The model has reclaimed the top position among publicly available models on key industry benchmarks. It achieved a score of 64.3% on both SWE-bench Pro and SWE-bench Verified, tests designed to evaluate a model's ability to handle complex software engineering challenges. This demonstrates a notable improvement in its capacity for multi-step, logic-based tasks.

Beyond coding, Opus 4.7 shows improved performance in agentic computer use, which involves autonomously navigating an operating system to complete tasks. It also exhibits stronger results in graduate-level reasoning tests, indicating a more thorough and consistent application of knowledge across difficult professional domains.

Interestingly, the new model shows a slight decrease in performance on cybersecurity vulnerability reproduction benchmarks, scoring 73.1% compared to the 73.8% achieved by Opus 4.6. According to Anthropic, this is an intentional outcome resulting from the introduction of new safeguards designed to automatically detect and block requests related to high-risk cybersecurity activities.

Contextualizing Opus 4.7 and the Mythos Preview

The release of Claude Opus 4.7 occurs alongside continued discussion of Claude Mythos Preview, Anthropic's most powerful model, which remains in limited access due to its advanced capabilities. The company's own benchmarking data consistently shows Mythos outperforming all other major models, including Opus 4.7, across nearly every test category.

Anthropic has been transparent about this hierarchy, stating that Opus 4.7 is "less broadly capable" than Mythos Preview. The strategy appears to involve testing new safety measures on less advanced models like Opus 4.7 before potentially wider implementation. While Opus 4.7 offers state-of-the-art performance to the public, it is also framed as a stepping stone toward the next frontier of AI development represented by Mythos.

Claude Opus 4.7 is available immediately across all Claude products and via the company's API, with no price increase. It represents a tangible advancement for users requiring high-level coding and engineering assistance, even as it points toward a more powerful future that Anthropic is carefully curating.

2026-04-10

Zhiyuan Robotics Unveils GO-2: A Leap Forward in Embodied AI for Robots

Zhiyuan Robotics unveils GO-2, a new embodied foundation model designed to enhance robot's ability to execute precise actions based on logical reasoning, improving stability and reliability in real-world applications.