Gemini 4 Hydrogen
Fast everyday execution. Everyday coding, multimodal work and agents.
Which reasoning effort setting?
controls how long Gemini 4 Hydrogen thinks before answering. Token prices never change; higher settings spend more and raise task cost.
Start at Xhigh. The top setting is also the best value here. Rule: cheapest setting within 2 points of the model’s best score.
At Xhigh, Hydrogen thinks for about 96.9K tokens per task, then writes about 57.5K more as its answer: $3.16 per task, 10.8× Low, for +27.0 points.
Output rate: $7.50/1M at every effort level. Thinking tokens are billed as output.
| Effort | Index | Per task | Thinking | All output | × Low | Time |
|---|---|---|---|---|---|---|
| Low | 26.9 | $0.29 | 6.1K | 14.1K | 1.00× | 1.5 min |
| Medium | 41.2 | $0.80 | 20.6K | 39.4K | 2.73× | 4.1 min |
| High | 50.8 | $1.87 | 47.5K | 85K | 6.41× | 8.8 min |
| XhighStart here | 53.9 | $3.16 | 96.9K | 154K | 10.82× | 15.6 min |
Per task of the Artificial Analysis Intelligence Index: the average over its nine evaluations, counting input, cache, and output tokens at list prices. These are long, agent-style benchmark tasks, much longer than a chat message.
Where Gemini 4 Hydrogen sits on the frontier
Intelligence Index against one 100K-input, 10K-output request.