As Google prepares to roll out its new Gemini 4 “Argon” model to the public, company staff are testing better versions that could further narrow the gap with competitors.
Most recently, employees tried a new version of Gemini 4 internally named Carbon, according to documents and screenshots viewed by Business Insider.
The model – perhaps a new update of Argon, also known as “checkpoint” – was made available in recent days on Google’s internal coding platform, Jetski.
One employee told Business Insider that Carbon “feels like Opus 5.5” for coding, referring to Anthropic’s most powerful model for long-term agent coding. However, they warned that they would have to test it further.
Google’s internal tests show how quickly the company is evolving with its new generation of Gemini 4 models, announced at the end of September. It also highlights how competitive Google is in the fast-paced race for AI coding agents. Anthropic and OpenAI have moved ahead among developers with models designed for complex engineering work.
Early versions of Argon tested internally were generally well received by staff, according to conversations with employees and internal discussion threads viewed by Business Insider.
However, one employee said that early versions of Argon felt behind on some coding tasks, comparing the model to Claude Opus 5, an older version of Anthropic.
The latest updates could close this gap, but Google is not guaranteed to release them. An internal document shows that Google tested a series of Gemini 4 models, internally named Argon, Barium and Carbon.
Google’s internal codenames are generally not in sync with their public names. For example, one document states that the internal model “Barium-B” was chosen to be the one that will be known publicly as Argon. It could not be determined whether Carbon will be another update under the Argon name or a different model from the Gemini 4 family.
Google declined to comment.
Argon, Barium, Carbon…
Staff discussed the new updates in messaging channels seen by Business Insider.
“It’s comparable to Opus 5.5,” said one employee, referring to the latest update. “Carbon is really good!” another wrote.
“The new Gemini pro next model is pretty nice,” another person posted, referring to the “Next” label that is usually given to new Gemini updates internally.
Google said last month that Argon had achieved exceptional performance on certain coding and knowledge work benchmarks, including in areas such as law and finance.
The company also said Argon has a strong focus on defensive cybersecurity. It said it would first make the model available to partners in its Fairwind program, an initiative to test cyber vulnerabilities in new models, before a wider public rollout.
Google is under pressure to show that it can still compete at the frontier in terms of model capabilities. The company this week announced a new “universal” Gemini AI agent designed for the workplace, as the battle over agent assistants heats up between the biggest AI companies.
Do you have something to share? Contact this reporter by email at hlangley@businessinsider.com or Signal at 628-228-1836. Use a personal email address and non-work device; here is our guide to sharing information securely.
Gn bussni
