Claude Opus 4.7 became generally available on April 16, 2026, and remains the newest production Claude model through late May 2026. No additional generally available Claude model shipped during May itself. Opus 4.7 is offered across the Claude apps and API, Amazon Bedrock (AWS's managed AI service), Google Cloud Vertex AI, and Microsoft Foundry, at the same price as its predecessor: $5 per million input tokens and $25 per million output tokens.

The May 2026 model news centers on Mythos, a security-focused Claude variant. On May 25, 2026, Anthropic stated it intends to make Mythos-class models available through a general release in the near future, once it has built stronger safeguards. Until then, Mythos is restricted to Project Glasswing partners and is not selectable in any everyday product. During testing, Mythos found thousands of real vulnerabilities, including a 27-year-old OpenBSD bug and a 16-year-old FFmpeg flaw that prior tooling had missed.

Why Most Apps Can't Offer New Claude Models on Day One?

Most applications that wrap a Claude model ship with the model identifier hardcoded directly into the binary. When Anthropic releases or renames a model, that app cannot offer it until the maintainer notices the change, updates the code string, and ships a new build. Users wait on the developer's release schedule, not Anthropic's. This friction is exactly what people searching for "Claude May 2026" are circling: the model exists, but their tool still shows the old one.

This architectural limitation creates a gap between what Anthropic releases and what users can actually access. A developer might take days, weeks, or even months to update their application after a new model lands. During that window, users with access to the latest Claude through their own accounts cannot use it within the app they rely on.

How Can Apps Stay Current Without Constant Updates?

Some tools are built differently. Instead of baking model names into the code, they ask the Claude backend which models are available at runtime and display whatever the backend advertises. This approach means a new model appears in the picker the moment your account can resolve it, without requiring any app update.

• Dynamic Model Discovery: The app queries the Claude Code Agent Client Protocol (ACP) backend to fetch the live list of available models instead of relying on a frozen list compiled into the binary.
• Fallback Defaults: A small static list of models exists only so the picker is not empty before the backend responds; the moment the real list arrives, it replaces the defaults entirely.
• Legacy Migration: When users upgrade the app, older stored model preferences are automatically mapped forward to the current canonical model ID, so a preference from an older build still resolves to a real, current model instead of silently falling back to something else.

The same pattern applies to other AI backends. Tools using Google's Gemini, for example, deliberately use the un-versioned aliases "gemini-flash-latest" and "gemini-pro-latest" rather than hardcoded version numbers, so the picker auto-rolls when Google ships a new generation.

What Does This Mean for Your Claude Workflow?

For day-to-day coding, browser work, or desktop automation, the model you should expect to be running right now is Opus 4.7. If and when a Mythos-class model reaches general release, a tool that reads its model list from the backend will surface it the same way it surfaced every Opus before it: as a new entry, when your account can resolve it, without requiring you to reinstall anything.

Availability still follows your Claude account's entitlements. An app does not grant access to a model your plan cannot resolve; it simply stops getting in the way once your plan can. If you are on Claude Pro or Claude Max, you will see the models your subscription tier supports. If you are using a third-party app with its own API key, the app's billing account determines what you can access.

The practical implication is clear: if you are choosing between tools that wrap Claude, ask whether the tool reads its model list dynamically from the backend or whether it hardcodes model names. The first approach means you will have access to new Claude models as soon as Anthropic releases them and your account can resolve them. The second approach means you will wait for the developer to ship a release, which may lag Anthropic's timeline by days or weeks.