Anthropic to Release Claude Opus 4.7 and New AI Design Tools, Rumors Say This Week

AIbase Report | April 16, 2026

In recent months, Anthropic has been updating at a pace that's hard to keep up with—nearly one major update every two weeks since January, rolling out models, Claude Code, Cowork, Excel plugins, and more. Now, the news is heating up again: According to The Information's exclusive report, Claude Opus 4.7 and a brand-new AI design tool may be released this week. On the day the news broke, Adobe, Figma, and Wix stock prices dropped by over 2%.

Opus 4.7 leaked on Google Vertex AI

What’s Coming This Time

First, the model. Opus 4.7 is an incremental upgrade from Claude Opus 4.6, which was released in February. Version 4.6 introduced enhanced coding capabilities, improved agent task execution, and a 1 million token context window. The internal API references for Opus 4.7 have recently surfaced, indicating that the release is near.

Developers in the community have already found switches for Opus 4.7 in API configurations and glimpsed its presence in the list of models on Google Vertex AI. The source code for Claude Code was also accidentally leaked, revealing references to Opus 4.7 and Sonnet 4.8, as well as the newly launched Routines feature—most of which were later officially fulfilled.

What really caught people's attention was the design tool. Anthropic is preparing to launch an AI-driven design tool that allows users to generate websites, presentations, and landing pages through natural language prompts. This tool targets both developers and non-technical users, directly competing with startups like Gamma and Google's Stitch. In other words, Anthropic no longer just wants to sell models—it aims to turn the idea of "inputting a sentence and getting a usable product" into a real product itself.

Why Did Adobe and Figma Drop?

There's a bigger context here. The S&P 500 software and services index has fallen nearly 26% this year, with investors worried that AI tools might reduce demand for traditional software products. When Anthropic previously launched the Claude Cowork assistant and a series of automation plugins, software stocks also experienced significant sell-offs.

To put it simply, the market isn't worried about one tool replacing another. It's concerned that once the "prompt-as-design" paradigm takes off, a large portion of the traditional toolchain could be disrupted. Anthropic has already partnered with Figma—converting AI-generated code into editable design files—and Claude is also integrated into Microsoft Word and PowerPoint. This "both collaboration and direct competition" approach makes things more uncertain than just launching a competitor.

Opus 4.7 Isn’t Anthropic’s Strongest Card

There's a detail worth highlighting in this report. Opus 4.7 is not the strongest model Anthropic has. That position belongs to Claude Mythos, currently being used by a few partners for security vulnerability research testing.

Mythos takes an entirely different path—Anthropic builds its cutting-edge models by fine-tuning the Opus series. Opus 4.7 is essentially the base that will eventually be "infused with cybersecurity skills" and hardened into Mythos. When the UK AI Safety Institute evaluated the Mythos Preview, they found it could autonomously execute complex network attacks that other models haven't reached yet. Therefore, Anthropic has kept Mythos locked within the Project Glasswing project, and it's expected to be more widely revealed in San Francisco in May.

This "dual-track" strategy actually reflects Anthropic's current stance: the commercial line (Opus) is constantly iterating and competing aggressively, while the cutting-edge line (Mythos) is focused on risk control behind closed doors, not rushing to release it.

By the Way: From $80 Billion to $800 Billion

Investors have valued Anthropic as high as $800 billion, more than doubling from the $380 billion valuation in the February funding round. On the secondary market platform Caplight, Anthropic's current transaction valuation is $688 billion, up 75% in three months. For comparison, OpenAI's latest valuation was $85.2 billion.

The surge in popularity is driven by revenue data. Anthropic claims its annualized revenue has risen from $9 billion at the end of 2025 to $30 billion, with over 1,000 enterprise customers spending over $1 million annually. This number has doubled in less than two months.

However, the cost is also becoming apparent. As compute costs soar, Anthropic is replacing fixed enterprise pricing with usage-based billing, which could double or triple costs for heavy users. This shift comes amid the increased reasoning costs caused by tools like Claude Code and Claude Cowork, and the tightening availability of compute resources.

For developers, this is a signal to plan ahead—your product's cost structure may need to be re-evaluated if it heavily relies on Claude Code or APIs.

This Week Will Be Busier Than Ever

This week has been called one of the busiest in AI history by many in the industry. In addition to Anthropic, OpenAI is also expected to release new updates, and Meta's LlamaCon is happening around the same time. This intense release schedule can easily overwhelm users, but it also means the window for choosing tools will be shorter than ever before.

A Quick Judgment

From Claude Code Routines to Opus 4.7 and the design tool, Anthropic's activities over the past six months have been clear: it's not building single-point AI capabilities, but a complete stack from models to tools to distribution. The design tool is a key step in this stack, as it brings non-technical users into the fold for the first time—before, Claude Code, Cowork, and Excel plugins mainly targeted developers and knowledge workers.

Several points to watch in the future:

• Pricing. The previous API price for the Opus series was $5/$25 per million tokens. Whether 4.7 follows this range will directly affect its value proposition compared to GPT and Gemini.
• Export of the design tool. Whether it becomes an independent product, integrates into claude.ai, or opens an API through the Claude Platform will have completely different ecosystem impacts.
• Boundaries of the Figma partnership. Working together while creating a competitor, how the product details are divided needs to be carefully observed later.

Officially, there is still no exact release date. The information above mainly comes from The Information's exclusive report, leaks in the Claude Code source code, API traces on platforms like Vertex AI, and community discussions. Given Anthropic's recent pace, the answer is likely to be revealed in a few days.