Gemini 4 Carbon early test results are in ahead of release (www.testingcatalog.com)

Google is ramping up its AI efforts with the early testing of Gemini 4 Carbon, the successor to the recently launched Gemini 4 Argon. Reports indicate that a select group of partners has begun accessing the new model via the LM Arena, where initial impressions suggest it shows significant improvements in 3D tasks, although it lags behind in frontend and UI capabilities and is slower overall. Evaluations from early access testers note Carbon’s outputs are of higher quality compared to Argon, though its speed varies from two to four times slower depending on the task at hand. The significance of Carbon's early access lies in its potential to refine Google's AI offerings amid increasing competition from other tech giants like OpenAI and Anthropic, both of whom are gearing up to release new models. As Google continues to process massive amounts of data—25 trillion tokens in just the first 72 hours following Argon’s launch—it aims to keep its momentum going. The development of Carbon, alongside improvements in Gemini's desktop app capabilities and upcoming models like Nano Banana 3 Pro, signals Google's strategic push to address existing weaknesses and maintain its competitive edge in the rapidly evolving AI landscape.