Google launches Gemini 4 Carbon, its strongest model
Gemini 4 Carbon is now available to paying developers via API and is rolling out to AI Ultra subscribers in Antigravity, the Gemini app, and Search. Google claims it ranks first on CursorBench 4.0, WANDR, and AutomationBench at more than 2x lower cost per task than the next best model, priced at $2/$10 per million tokens.
Why it matters
๐ป Developer ยท Same pricing as Argon but claimed top benchmark scores on coding and automation tasks, worth testing against your current model.
๐ฆ Product ยท A cheaper top-tier model could cut agent and automation costs if Google's benchmark claims hold up in your use case.
๐จ Design ยท Not a design-focused release, but powers the Antigravity tools designers may already be using for prototyping.
๐ Business ยท Lower cost-per-task at frontier performance is a real lever if you are running AI at scale, though claims are Google's own.
๐ค Just Curious ยท Google's newest flagship AI model is out, and it's reportedly both smarter and cheaper to run than its predecessor.
Sources: Gemini 4 Carbon is now available, Gemini 4 Carbon tool page