Gemini 4 Hydrogen ranks fifth in Agent Arena and tenth in WebDev
Arena reports that Google DeepMind’s Gemini 4 Hydrogen (High) ranked fifth overall in Agent Arena across more than 6,000 real-world agentic sessions, rising to fourth on knowledge-work tasks. It also debuted tenth in Code Arena: WebDev with 1,703 points and placed fifth in HTML and Content Creation Tools. Arena reported a median task time of 5.3 minutes and API pricing of $1.50 per million input tokens and $7.50 per million output tokens.
The results position Gemini 4 Hydrogen as a fast, cost-efficient option for agentic and knowledge work, while also placing it among leading models in web-development benchmarks.
Gemini 4 Hydrogen excels at knowledge work.
Based on over 6,000 real-world agentic sessions, Hydrogen (High) is #5 in Agent Arena overall. On knowledge-work tasks, it moves up to #4.
As a result, it’s ahead of three rival models that outrank it overall: Claude Opus 5.5 (Max), GPT-6 Astra (Max) and Claude Sonnet 5.5 (Max).
This means that @GoogleDeepMind's Hydrogen is not only the fastest model in the Agent Arena top 5, but also a powerful choice for real-world work.
