Google has begun releasing its newest artificial intelligence model, Gemini 4 Carbon, to the general public. This rollout follows a two-week period during which the model was exclusively available to a select group of cybersecurity professionals.
The tech giant asserts that Carbon represents its most powerful system to date, positioning it as a direct competitor to the leading AI models from Anthropic and OpenAI. According to Google, the new model significantly advances the frontier beyond Argon, the company’s other frontier model that launched just last month.
While Google suggests that Argon will remain the preferred option for most software engineering and knowledge work tasks, Carbon is specifically designed for applications that push the boundaries of a model’s intelligence.
Curious how it handles creative writing. Previous Google models were decent but often lacked nuance.
Carbon sounds flammable. I hope this model does not cause any major data center fires.
The public rollout is fast. Last month Argon was still in internal testing phases apparently.
Can we get pricing details soon? I need to know if this is free or another subscription tier.
Keeping Argon for coding makes sense. Specializing models is becoming the standard industry practice now.
Rivalling OpenAI is bold. We will see actual benchmarks before believing the marketing hype.
Two weeks for cybersecurity pros first? Feels like a security patch disguised as an AI release.
Surprised they named it Carbon. Feels weird to have a heavy model named after a light element.