Anthropic officially launched Claude Opus 4.7 on April 16, 2026, presenting it as its most capable broadly available AI model to date. The release builds on Opus 4.6, which arrived only two months earlier, with upgrades aimed at software development, long-horizon task execution, and advanced financial analysis. The move also reflects a wider industry shift toward agentic AI systems that can handle more autonomous workflows.
Stronger results in coding and finance
According to Anthropic, Claude Opus 4.7 reached 87.6% on SWE-bench Verified, up from 80.8% for Opus 4.6. The company also said the new model outperformed OpenAI’s GPT-5.4 and Google’s Gemini 3.1 Pro in tests related to tool use and computer interaction, two areas that are increasingly important for autonomous software agents.
In financial analysis, the model posted a 0.813 score in the General Finance module, compared with 0.767 for the prior version. Anthropic said this points to improved rigor in economic modeling and more reliable performance in complex finance-related tasks, making the model more appealing for professional use cases.
Vision upgrades and multi-agent tooling
One of the clearest changes is in visual processing. Claude Opus 4.7 can now handle images up to 2,576 pixels on the long edge, roughly triple the previous limit. That upgrade is designed to improve interpretation of detailed charts, technical diagrams, and user interface screenshots, though Anthropic noted that higher-resolution images also consume more tokens and may raise costs.
The company also introduced a new /ultrareview feature inside Claude Code, allowing professional and Max-tier users to run multi-agent sessions to identify bugs and design flaws in software. The model remains available through Claude API, Amazon Bedrock, Google Vertex AI, and Microsoft Foundry, while keeping the 1 million-token context window introduced earlier in the year.
Higher capability, but token efficiency matters
Anthropic kept pricing unchanged at $5 per million input tokens and $25 per million output tokens. Even so, some testers observed that the updated tokenizer may consume up to 1.35 times more tokens for the same input. That could affect developers running long autonomous tasks, so Anthropic has added a task budgeting feature in public beta to help manage spend.
Early feedback from developers suggests Claude Opus 4.7 follows instructions more literally than earlier versions, which may require prompt adjustments for existing workflows. Anthropic also said safety remains central to the release, with new automated guardrails intended to block high-risk cybersecurity use cases and strengthen resistance to harmful output generation.
Overall, Claude Opus 4.7 appears to be more than a routine model refresh. It underscores how competition in AI is moving beyond chatbot performance and toward autonomous execution, tool use, and practical deployment in specialized professional environments.