Anthropic Introduces New Opus 4.7 Model Focused on Advanced Software Engineering

Publication Date: 16.04.2026

Rate this article:

4.8/5 ( 31 votes )

Table of Contents:

Anthropic has announced its latest artificial intelligence model, Claude Opus 4.7. The new version comes two months after the update of the previous model and aligns with Anthropic's previous update frequency.

Opus 4.7 Requires Less Supervision for More Challenging Coding Tasks

Claude Opus 4.7 is Anthropic's latest generally available version focused on advanced software development.

Opus 4.7 offers a significant improvement over Opus 4.6 in advanced software engineering, providing notable gains especially in the most challenging tasks. Users report that they can confidently delegate the most difficult coding jobs, which previously required close supervision, to Opus 4.7. Opus 4.7 manages complex and long-term tasks meticulously and consistently, paying precise attention to instructions and developing ways to validate its own output before providing feedback.

Anthropic notes that the model has a better vision and enjoys creating higher quality work.

The model also has a significantly better vision: it can see images in higher resolution. It produces more enjoyable and creative, higher-quality interfaces, slides, and documents while completing professional tasks.

The company presents positive comparisons in various use cases, showing that Opus 4.7 is ahead of 4.6, GPT-5.4, and Gemini 3.1 Pro, but falls behind Claude Mythos Preview, which has broader capabilities. However, Mythos is not generally available as it is only shared with major software platform vendors like Apple.

You can view the comparison chart in Anthropic's blog post here.

Anthropic highlights other improvements in Opus 4.7, such as instruction tracking, multimodal support, real-world tasks, and memory.

“Opus 4.7 is better at using file system-based memory,” says the company. “It remembers important notes across long, multi-session works and uses them to transition to new tasks; thus, it requires less prior knowledge.”

Anthropic Sends Claude Opus Updates Every Two Months

Notably, Anthropic has established a more predictable frequency for directly updating the Claude Opus model.

Opus 4.7 arrives two months after Opus 4.6; Opus 4.6 came two months after Opus 4.5. There was a three-month gap between Opus 4.1 and Opus 4.5.

Anthropic's announcement includes a note for users on how token usage will be managed with Optus 4.7:

Opus 4.7 is a direct update to Opus 4.6, but there are two changes worth noting as they affect token usage. First, Opus 4.7 uses an updated tokenizer that improves the model's text processing. As a result, the same input may correspond to more tokens—approximately 1.0–1.35× depending on the content type. Second, Opus 4.7 considers higher effort levels in subsequent stages, particularly in agency settings. This increases its reliability on difficult problems but means it generates more output tokens.

The company has a separate post regarding the transition, and the Claude Opus 4.7 System Card is accessible here.

In addition to new models, Anthropic has also been working on Claude Code, part of the Claude Mac application, in recent weeks:

Anthropic is giving Claude the ability to use your Mac for you
Anthropic is adding routines to the redesigned Claude Code, here's how it works
Claude Code offers a safer alternative for developers to skip permissions called ‘auto mode’