
AI and SaaS Roundup · 2026-11-12
Google’s Carbon model is available more widely, Claude can work with you in live sheets and whiteboards, and Qwen4 has a new preview. There are a couple of dates to know before you build around these. 👀
Gemini 4 Carbon is now available to paying developers
Google opened Gemini 4 Carbon more widely on Tuesday, after its earlier limited cybersecurity release. Developers can use the paid Gemini API through AI Studio; Google AI Ultra subscribers also get access through products including Antigravity and the Gemini app.
It’s aimed at harder, longer jobs such as multi-step debugging, architectural changes and working through complex files. Google also previewed /swarm in Antigravity for coordinating agents across a larger project.
The introductory API price is $2 per million input tokens and $10 per million output tokens. That introductory period ends on 30 November 2026, after which Google lists $4 input and $20 output. So the launch price isn’t the ongoing rate.
Claude adds editable Sheets and Whiteboards
Claude Sheets and Whiteboards launched in beta on Tuesday, while Claude Dashboards became generally available as Claude Data. They’re available on paid plans.
Sheets lets you work with Claude on spreadsheet data, formulas and multiple tabs. Whiteboards gives you a shared space to lay out ideas and ask Claude to organise or summarise them. For a builder, that could be a launch plan on a board followed by a pricing or campaign spreadsheet you can keep editing.
These work on web, desktop and mobile. Team plans have all three enabled by default. Enterprise is different: Sheets and Whiteboards start off disabled, while Data turns on by default on 19 November unless the admin changes the settings. Data needs a connected source such as a warehouse or CRM.
Qwen4-Max-Preview is live, but this version has an expiry date
Alibaba’s Qwen4-Max-Preview arrived on Wednesday. It accepts text, images and video, with a 1M-token context window, tool calling and structured output. It’s pitched for coding, frontend work and longer agent tasks.
Access is through Alibaba Cloud’s Token Plan Personal or Team, using a dedicated key and endpoint. Pay-as-you-go access isn’t supported for this preview. The model ID is qwen4-max-preview, with support for OpenAI and Anthropic API formats.
The preview runs until the end of 25 November 2026 in UTC+8, and Alibaba says this model will be retired afterwards. Treat it as something to evaluate during that window. One useful implementation detail: its max_tokens budget covers both thinking and the final answer, so an answer-only budget can cut the response short.

Comments
Sign in to join the conversation
Sign InNo comments yet. Be the first to share your thoughts!