Key points
- Google staff are testing a model called Carbon
- One employee compares it to Anthropic's coding model
- Argon still hasn't reached the public
Google has begun testing Carbon, an internal Gemini 4 model that one employee said "feels like Opus 5.5" for coding, Business Insider reported on Oct. 9. The employee cautioned that it needed more testing.
Carbon became available in recent days through Jetski, Google's internal coding platform. The comparison is to Anthropic's Claude Opus 5.5, which Business Insider described as Anthropic's most powerful model for long-running coding tasks. Another employee wrote in an internal channel: "Carbon is really good!"
Google declined to comment. Business Insider could not establish whether Carbon would become an Argon update or a separate Gemini 4 model. Google may never release it publicly.
Argon has not reached the public yet
Google unveiled Gemini 4 Argon on Sept. 30 and gave first access to cyber defenders in its Fairwind Program, as we reported at the time. Google said paid API customers and Google AI Ultra subscribers would get access next, but it has not given a date.
Early internal versions of Argon were generally well received by staff, according to Business Insider. One employee said those versions lagged on some coding tasks and compared them to Claude Opus 5, an older Anthropic model.
Google's internal codenames often don't match its public model names. One internal document said a model called "Barium-B" was chosen to become the public Argon, according to Business Insider.
Google's own benchmark table put Argon first on 13 of 19 results, while independent leaderboards gave it a more mixed ranking. Business Insider reported that Anthropic and OpenAI have pulled ahead with developers on models built for complex engineering work.
Alphabet (GOOGL) shares were up 1% at $351.60 at 3:54 p.m. Eastern on Oct. 9.
