Latest News About Opus 4.7 Xhigh

Here’s what’s latest on Opus 4.7 xhigh, based on recent reports:

• Claude Opus 4.7, Anthropic’s flagship model, was released in mid-April 2026 and is presented as a notable upgrade over Opus 4.6, with improvements in software engineering, reasoning depth, and vision capabilities. It also introduces a new mid-to-high reasoning setting called xhigh, aimed at balancing depth of reasoning with latency on harder tasks.[1][2][3]

• Key new capabilities include: support for higher-resolution vision inputs (up to about 2,576 pixels on the long edge, roughly 3.75 megapixels), improved performance on complex multi-step tasks, and stronger instruction following. Pricing remains unchanged from Opus 4.6 at $5 per million input tokens and $25 per million output tokens.[2][1]

• There is some mixed coverage about practical gains: several outlets highlight stronger coding, document analysis, and long-running task handling, while others note that the magnitude of gains can vary by workload and that some early community commentary questions whether the new xhigh tier delivers universal improvements.[5][7][1]

• Anthropic also signaled ongoing safety and cybersecurity safeguards tied to related initiatives, with Mythos Preview existing behind closed doors and Opus 4.7 serving as the broadly available model with built-in safeguards.[3][2]

• Public reception and analysis vary: tech summaries and industry chatter discuss the tradeoffs of the new xhigh setting, practical benefits in coding and engineering tasks, and ongoing comparisons to Mythos Preview. Some coverage even notes user experiences may change as prompts are aligned more literally to instructions.[4][7][3]

If you’d like, I can pull the most recent pundit opinions or summarize developer-facing release notes in a quick bullet list, and I can also compare Opus 4.7 to Opus 4.6 side-by-side. I can also look for any updates on availability across platforms (e.g., Bedrock, Vertex AI, Foundry) and any early benchmarks you care about.[1][2]