Anthropic launches Opus 4.7, a Claude‑powered rival to Figma, as Schneier warns on

Anthropic’s Claude Design runs on the newly announced Claude Opus 4.7 vision model, the company’s most capable multimodal system, according to the official launch post on elsolitario.org. The tool lets users request full‑featured design assets through a conversational interface, producing editable documents with components, typography, grids and interactive states rather than static PNGs. Anthropic says the service is immediately available in research preview for Pro, Max, Team and Enterprise plans via claude.ai/design.

The move marks Anthropic Labs’ first consumer‑facing product beyond chat, stepping into a market worth over $25 billion annually in SaaS design and content production, the same source notes. By generating wireframes, interactive prototypes and design explorations on demand, Claude Design aims to undercut incumbents such as Figma, Canva and Adobe XD, which rely on manual drag‑and‑drop workflows.

Anthropic’s rollout coincides with heightened scrutiny of its security research. Bruce Schneier warned last week that the company’s Claude Mythos model, capable of discovering and weaponising software vulnerabilities, is being kept off‑limits to the public and shared only with roughly 50 critical‑infrastructure partners under “Project Glasswing,” per Schneier’s commentary. He highlighted that Mythos uncovered thousands of bugs, including a 27‑year‑old OpenBSD flaw and a 16‑year‑old FFmpeg issue, and generated 181 usable Firefox attacks versus two from Anthropic’s prior model.

Schneier cautioned that while Anthropic reported an 89 percent severity‑rating agreement with security contractors on 198 flagged issues, the lack of full disclosure makes it impossible to gauge false‑positive rates. Independent research shows AI bug‑finders often hallucinate vulnerabilities in patched code, a risk that could amplify if models like Mythos are later integrated into broader products such as Claude Design.

The juxtaposition of a design‑focused launch and a tightly controlled security model underscores Anthropic’s dual strategy: expand into high‑value creative SaaS while containing potentially disruptive AI capabilities. Observers will watch whether the research preview gains traction among enterprise designers and how the company balances rapid product rollout with the responsible‑disclosure concerns raised by Schneier.