Sakhanda Wire
NVDA $154.51 -1.04% MSFT $506.49 -2.22% GOOGL $281.27 -0.86% META $621.51 -0.78% AMZN $256.90 -0.33%
← Back to the news

Gemini 4 Carbon becomes generally available as model competition accelerates

Google has made Gemini 4 Carbon generally available to paying developers and business customers after a two-week early-access period for selected security experts. The model is positioned alongside Gemini 4 Argon, not as its replacement: Google describes Argon as the better option for most software engineering and knowledge-work tasks, while Carbon targets demanding multi-step coding, debugging and open-ended scientific research. Carbon uses more tokens for deeper reasoning and verification, so its total cost can exceed Argon’s despite identical token prices. Its introductory price is $2 per million input tokens and $10 per million output tokens.

Google says Carbon leads several benchmarks, including CursorBench 4.0, Perplexity’s WANDR and Zapier’s AutomationBench, while costing less than half as much per task as the next-best model, Fable 5.2. In the Artificial Analysis test, Carbon averaged $7.76 per solved task. It placed second on the Multimodal Knowledge Agent benchmark, behind Fable 5.2.

Google is also previewing /swarm in Antigravity, which coordinates multiple agents working in parallel. The company additionally reports security gains, including zero successful prompt injections in one internal benchmark and a detection-rate improvement of more than 25 percent over Argon in Project Fortify.

Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.

Originally published by The Decoder on

Read the original on The Decoder ↗

The reporting and the image belong to The Decoder and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news