Google opens Gemini 4 Carbon at Argon's introductory API price, doubling after November 30, and Carbon ranks second on a third-party-run knowledge-work benchmark

Concept illustration in two side-by-side panels showing the same scene: an open laptop whose dark screen shows a speech bubble with three lines, and above it two labels, one lettered GOOGLE beside a multicoloured G and one lettered GEMINI 4 CARBON, each joined by a thin line to the laptop's top edge. In both panels a stack of red coins stands to the laptop's left, short in the left panel and roughly twice as tall in the right panel. Printed on the image: US$2 → US$4 Carbon API price per million input tokens, introductory vs after November 30, per Google.

Google DeepMind opened its new model Gemini 4 Carbon on November 10 (Google DeepMind, 2026-11-10). Google says Carbon costs under one-sixth as much per task as Anthropic's top-ranked Claude Fable 5.2, without saying whether that uses the introductory price or the doubled price after November 30.

Google positions Carbon for the most demanding tasks and says Argon, Google's current main model, remains the first choice for most software engineering and knowledge work. Carbon's introductory price is US$2 per million input tokens and US$10 per million output tokens, the same as Argon's introductory price. After November 30, Carbon moves to US$4 and US$20.

The third-party figures come from MKA, a knowledge-work agent benchmark designed by Google DeepMind and run independently by Artificial Analysis, a site that evaluates AI models. Anthropic's Claude Fable 5.2 at its highest reasoning-effort setting, Max, ranks first at 57%; Carbon at High scores 54%, and Claude Opus 5.5 at Max scores 53% (Artificial Analysis). Scores and cost per task both change with the effort setting. Google says Carbon's average cost per task on MKA is under one-sixth of top-ranked Fable 5.2's. That comparison comes from Google's release post.

The other results are Google's own. On a 375-question agent honesty benchmark Google built, Carbon passes 92% and Argon 70%.

Not proven yet: the release post doesn't say whether the cost comparison uses the introductory or doubled price How we checked →

Next to watch: cost per task as recalculated by third parties such as Artificial Analysis after the introductory price ends on November 30.

Source and checks

This item ran in the brief of Nov 11, 2026, alongside 15 other items. How we checked it, the quote checks and what was still missing that day are on that issue's How we checked page.

Cite this item

SecondSource (2026). "Google opens Gemini 4 Carbon at Argon's introductory API price, doubling after November 30, and Carbon ranks second on a third-party-run knowledge-work benchmark." https://secondsource.io/en/news/2026-11-11-3/. Accessed access date.