Model
Gemini 4 Carbon.
Power
££
API cost
In plain English
Google's strongest current model.
Good for
Complex coding, scientific research, cybersecurity defence and long-running agents.
Things to be aware of
It costs more and runs more slowly than lighter models. Keep it for work where the extra ability changes the result.
What Google DeepMind says
“Carbon is built for tasks that “test the limits of a model’s intelligence”.”
What everyone else says
Independent rankings change by task. Carbon leads some agentic benchmarks, while lighter models finish other jobs faster and at a lower cost. Artificial Analysis
- Context window
- 1.0M tokens
- Knows up to
- March 2026
- Run it yourself
- No, online only