Anthropic expands product surface
- Anthropic launched Claude Opus 4.7 on April 16 and Claude Design on April 17, widening Claude from a coding model into a design-and-prototyping product for paid users. - Opus 4.7 keeps Opus 4.6 pricing at $5 per million input tokens and $25 per million output tokens, while adding 1 million-token context, 128,000-token output, high-res vision, and new task budgets. - The expansion also adds migration risk: Opus 4.7 removes extended thinking parameters and Claude Design is still a research preview with separate limits and billing. (platform.claude.com) (support.claude.com)
Anthropic spent two days in mid-April widening Claude’s job description: first a new flagship model for coding, then a design tool built on top of it. (anthropic.com 1) (anthropic.com 2) On April 16, Anthropic made Claude Opus 4.7 generally available across Claude, its application programming interface, Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry. The company said the model improves advanced software engineering, long-running coding tasks, and vision. (anthropic.com) (support.claude.com) Opus 4.7 keeps the same list price as Opus 4.6 at $5 per million input tokens and $25 per million output tokens. Anthropic’s docs say it supports a 1 million-token context window, up to 128,000 output tokens, adaptive thinking, and the same tool set as Opus 4.6. (anthropic.com) (platform.claude.com) The model’s biggest product additions are around seeing and planning. Anthropic says Opus 4.7 is its first Claude model with high-resolution image support, raising the maximum image size to 2,576 pixels or 3.75 megapixels from 1,568 pixels or 1.15 megapixels. (platform.claude.com) Anthropic also added a new “xhigh” effort level and a beta feature called task budgets, which give the model a token countdown for an entire agent loop. The company says that helps Claude decide how much work to do before it runs out of budget. (platform.claude.com) Then, on April 17, Anthropic launched Claude Design in research preview for Pro, Max, Team, and Enterprise subscribers. Anthropic describes it as a chat-and-canvas tool for making designs, prototypes, presentations, and one-pagers. (anthropic.com) (support.claude.com) Claude Design is powered by Opus 4.7 and can pull in screenshots, documents, slide decks, and code repositories. Anthropic says teams can also set up a design system so projects inherit brand colors, typography, and reusable components automatically. (anthropic.com) (support.claude.com) The expansion comes with caveats for developers already on Opus 4.6. Anthropic’s migration guide says Opus 4.7 removes the old extended thinking setting, so requests using `thinking: {type: "enabled", budget_tokens: N}` will return a 400 error. (platform.claude.com) Anthropic also says Opus 4.7 is “more direct and opinionated” than Opus 4.6, with less validation-forward phrasing and fewer emoji. Teams that tuned prompts around tone or old reasoning controls are being told to re-test before swapping model names. (platform.claude.com) Claude Design has its own operational wrinkles too. Anthropic’s help docs say the product is metered separately from chat and Claude Code, has weekly allowances by plan, and does not yet support audit logs or usage tracking. (support.claude.com) For Anthropic, the April releases push Claude further beyond a single chatbot or coding assistant. The company is now selling one stack for code, agents, office work, and visual design — with separate docs, controls, and billing for each layer. (support.claude.com) (platform.claude.com)
