Google's Gemini 4 "Carbon" model reportedly feels like Anthropic's Opus 5.5 coding performance

Google is testing multiple variants of its upcoming Gemini 4 model family internally, according to documents, screenshots and internal chats seen by Business Insider. The variants are named Argon, Barium and Carbon. Carbon was deployed on Google's internal coding platform Jetski over the past few days and is said to outperform Argon mainly on programming tasks. One employee compared Carbon to Anthropic's strongest coding model, Opus 5.5, though the report notes it still needs more testing. Early Argon versions reminded another employee of the older Opus 5 on some coding tasks, even though internal feedback was positive overall.

Argon is Google's frontier reasoning model. Early rumors had pegged Argon as a Flash model optimized for speed rather than peak performance, but the announcement of Google's Gemini agent for Google Workspace put that to rest. Google describes its model lineup by role: Argon for frontier reasoning, Flash for speed and volume, Omni for generative media, and Gemma for lightweight, open-weights edge workloads. That places Argon as Google's most capable reasoning model, comparable to Anthropic's Opus or OpenAI's Astra. Carbon and Barium are likely checkpoints or updates within the Argon family rather than separate tiers.

One employee internally called Carbon "Gemini pro next model," suggesting it would ship under the Argon name. That fits with internal documents showing the now-unveiled Argon was previously called "Barium-B." Whether Carbon ships as an Argon update or a standalone model remains unclear, according to the report.

The quick update pace, similar to what Google has done with its Flash models, suggests Google has gotten better at using AI to build new models. Google Deepmind employee Vedant Misra responded to the Business Insider report on X by writing, "Have you heard of recursive self improvement," referring to using AI to build better AI. OpenAI and Anthropic have reported similar progress.

There is still no official Gemini 4 launch date, but Google is already retooling its apps. The Gemini app now shows a new "Automatic" mode for the 3-series models and adjustable reasoning intensity from low to high. Google AI Studio has surfaced a new "Ultra" mode promising "advanced skills and tools." Google's Logan Kilpatrick confirmed the team is working on getting the most out of Argon.

Why it matters: These reports point to a fast-moving internal cadence at Google, and if Carbon's coding performance really lands near Opus 5.5, teams building on frontier coding models could see a new competitive option in the near term. For developers, the practical signal is to watch for Argon-family releases and the new reasoning-intensity controls in the Gemini app and AI Studio, since those settings suggest Google is tuning models for different latency and quality trade-offs. It is worth treating the Opus 5.5 comparison as an unverified internal impression until benchmarks or an official launch confirm it. The naming uncertainty around Carbon, as an Argon update or a standalone model, means product teams should not yet plan around a specific release name or date.