PANews posted on X (formerly Twitter) that Anthropic has launched its latest model, Claude Opus 4.7, which focuses on improving performance in complex software engineering and long-process tasks. According to the company, this model surpasses its predecessor, Opus 4.6, in areas such as automatic planning, identifying and correcting logical errors, and handling complex asynchronous and CI/CD workflows. In tests conducted with multiple partners, the task completion rate improved by approximately 13% across 93 coding assessments. Additionally, the model has demonstrated the ability to solve more real-world production tasks on benchmarks like Rakuten-SWE-Bench.
Claude Opus 4.7 also enhances high-resolution image understanding, capable of processing images up to 3.75 million pixels. This feature is particularly useful in multimodal scenarios such as code review, document analysis, and life sciences patents. Despite these advancements, the pricing remains consistent with Opus 4.6. The API model is labeled as 'claude-opus-4-7,' and it introduces new control options like 'xhigh' inference intensity and task budgeting.