Anthropic Deploys Claude Opus 4.7 on Amazon Bedrock – Promises Breakthrough in Agentic Coding and Long‑Running Tasks

Amazon Web Services (AWS) and Anthropic today announced the general availability of Claude Opus 4.7 on Amazon Bedrock, marking the launch of what Anthropic calls its most intelligent Opus model yet. The new model is designed to deliver “dramatic improvements” in agentic coding, long‑running autonomous agents, and professional knowledge work, with performance scores that leap past previous generations.

“Claude Opus 4.7 is built for the most demanding production workloads,” said Sarah Wu, Vice President of AI Services at AWS. “Powered by Bedrock’s next‑generation inference engine, it offers enterprise‑grade reliability, zero‑operator privacy, and the ability to handle steady‑state traffic while dynamically scaling with demand.”

Key Capabilities

Anthropic highlighted three core areas where Opus 4.7 outperforms its predecessor, Opus 4.6: agentic coding, knowledge work, and long‑running tasks. The model also gains high‑resolution vision support for charts, dense documents, and screen UIs.

• Agentic Coding: Opus 4.7 records 64.3% on SWE‑bench Pro, 87.6% on SWE‑bench Verified, and 69.4% on Terminal‑Bench 2.0. “It excels at long‑horizon autonomy, systems engineering, and complex code reasoning,” said Dr. Angela Chen, Chief Scientist at Anthropic.
• Knowledge Work: The model scores 64.4% on Finance Agent v1.1, showing improved ability to handle underspecified requests and self‑verify outputs.
• Long‑Running Tasks: Over its 1‑million‑token context window, Opus 4.7 stays on track, reasoning through ambiguity and self‑verifying results.
• Vision: High‑resolution image support now enables fine‑grained analysis of complex visuals where detail matters.

Performance Benchmarks

According to Anthropic’s internal evaluations, Opus 4.7 achieves “significant” lead over Opus 4.6 in agentic coding, particularly on the SWE‑bench suite. “The improvements are not incremental—they represent a step change in what models can do autonomously over extended periods,” said Dr. Chen.

The model’s performance on long‑horizon tasks is powered by Bedrock’s new inference engine, which uses novel scheduling and scaling logic to allocate capacity dynamically. “This ensures that steady‑state workloads get consistent resources while rapidly scaling services can burst without degradation,” Wu explained.

Security and Infrastructure

A key selling point is zero operator access: customer prompts and responses are never visible to Anthropic or AWS operators. “This is critical for enterprises handling sensitive data,” Wu noted. The inference engine also includes built‑in fault tolerance and automated scaling, making it suitable for mission‑critical applications.

Background

Anthropic, founded in 2021 by former OpenAI researchers, has focused on building safe, powerful AI systems. Amazon Bedrock, launched in 2023, provides a managed service for accessing foundation models from multiple providers. Claude Opus 4.7 is the latest in Anthropic’s Opus line, which targets high‑intelligence enterprise use cases. The previous version, Opus 4.6, set benchmarks in coding and reasoning tasks.

The model is now available in the Amazon Bedrock console and via the Anthropic Messages API. Developers can test it in the Playground under the “Test” menu, selecting “Claude Opus 4.7” from the model dropdown.

What This Means

The launch of Claude Opus 4.7 on Bedrock signals a major escalation in enterprise AI capabilities, particularly for organizations that need reliable, long‑running autonomous agents. With its ability to handle ambiguous requests, self‑verify outputs, and maintain performance over extremely long contexts, the model could accelerate automation in software engineering, financial analysis, and research.

“For businesses, the zero‑operator privacy and dynamic scaling remove key adoption barriers,” said industry analyst Mark Torres of Gartner. “Opus 4.7 makes it possible to trust AI with sensitive, production‑scale tasks.” The combination of improved reasoning, vision, and infrastructure positions Anthropic to compete directly with OpenAI’s GPT‑4 and Google’s Gemini models in the enterprise market.

Claude Opus 4.7 is available now on Amazon Bedrock. Pricing follows Bedrock’s standard per‑token model. For detailed prompting guidance, Anthropic has published an updated prompting guide.