Claude Opus 4.7: Anthropic Releases Its Most Powerful Publicly Available AI Model

[Anthropic](https://anthropic.com) just raised the bar. On April 16, 2026, the company launched Claude Opus 4.7, the latest version of its flagship model. Smarter, faster, more precise — Opus 4.7 pushes the boundaries of autonomous coding, computer vision, and agentic workflows. And unlike Claude Mythos, its ultra-powerful sibling restricted to an elite consortium, Opus 4.7 is available to everyone starting today.

What's actually new in Opus 4.7

Opus 4.7 isn't a cosmetic update. It's a qualitative leap on the dimensions that matter most for developers and enterprises. Here are the key improvements:

• +13% on a 93-task coding benchmark — including 4 tasks that no previous Claude model (Opus 4.6, Sonnet 4.6) could solve.
• 3x more production tasks resolved on Rakuten-SWE-Bench, with double-digit gains in Code Quality and Test Quality.
• High-resolution vision up to 2,576 pixels on the long edge (~3.75 megapixels) — 3x more than previous models.
• New tokenizer that improves text processing (with a 1.0-1.35x impact on token count).
• Substantially better instruction following — Opus 4.7 interprets instructions literally, which may require re-tuning your prompts.
• Better memory — the model retains important notes better across long, multi-session work.

Benchmarks: Opus 4.7 vs Opus 4.6 vs GPT-5.4

The numbers speak for themselves. Opus 4.7 outperforms Opus 4.6 on virtually every benchmark, and competes directly with GPT-5.4 and Gemini 3.1 Pro:

A leap forward for autonomous coding

Coding is where Opus 4.7 shines brightest. Early-access partners are unanimous: the model handles complex, long-running tasks with newfound rigor. It verifies its own outputs before submitting them, catches its own logical faults during planning, and keeps executing through tool failures — where previous models would give up.

“Claude Opus 4.7 autonomously built a complete Rust text-to-speech engine from scratch — neural model, SIMD kernels, browser demo — then fed its own output through a speech recognizer to verify it matched the Python reference. Months of senior engineering, delivered autonomously.”

For Claude Code users, Anthropic has raised the default effort level to "xhigh" — a new tier between "high" and "max" that offers a better balance between reasoning depth and latency. The new /ultrareview command generates a dedicated review session that reads through changes and flags bugs like a careful human reviewer would.

High-resolution vision: a new playing field

Opus 4.7 sees better. With native support up to 3.75 megapixels (2,576 pixels on the long edge), the model can analyze dense screenshots, extract data from complex diagrams, and work with pixel-perfect visual references. This is a paradigm shift for agents using Computer Use — Claude's computer control feature.

XBOW, which specializes in autonomous penetration testing, reports a score of 98.5% on its visual acuity benchmark, up from just 54.5% for Opus 4.6. Solve Intelligence is already using the improved vision to read chemical structures and interpret complex technical diagrams for life science patent workflows.

Security and cyber defense: Mythos safeguards

Opus 4.7 arrives in a particular context. The week before, Anthropic unveiled Project Glasswing and its ultra-powerful Claude Mythos Preview model — capable of discovering thousands of zero-day vulnerabilities. Opus 4.7 is the first model to integrate new cyber safeguards developed for Mythos: automated detection systems that block requests related to prohibited or high-risk cybersecurity uses.

Pricing and availability

Good news: Opus 4.7 keeps the same pricing as Opus 4.6. No price increase for the upgrade.

What this means for Claude users

For developers, Opus 4.7 is a direct upgrade from Opus 4.6. Anthropic recommends re-tuning your prompts — the model follows instructions more literally, which may produce unexpected results with prompts written for earlier models. The new tokenizer may also slightly increase token counts for the same text (1.0-1.35x).

For consumers on Claude Pro or Max, the transition is seamless. Opus 4.7 replaces Opus 4.6 as the default model. It's more creative and more opinionated — it proposes solutions rather than simply agreeing, and produces higher-quality interfaces, presentations, and documents.

To learn more, check out our Claude vs ChatGPT comparison and discover the latest Claude features like Dispatch, which lets you control your desktop from your phone.

Frequently asked questions