Gemini 4 Hydrogen

gemini-4-hydrogen is the Gemini 4 id Google opened on November 12, 2026 for everyday coding and agentic work. The Gemini API model page names that string, a unified 1,048,576-token context, and a default synchronous output cap of 65,536 tokens, with long decode up to 1,048,576. The model card, published November 12, says Hydrogen is adapted from Gemini 4 Carbon. The knowledge cutoff on that card is March 2026.

The Gemini API pricing page, read November 13, 2026, prices the paid standard tier at $1.50 per million input tokens and $7.50 per million output tokens, including thinking tokens. The standard free tier on the same card lists input and output as free of charge, and the free-tier column is marked as used to improve Google’s products. Paid standard is marked no. Batch and Flex are $0.75 and $3.75 and are not on the free tier. Priority is $2.70 and $13.50. The card does not print a dated price increase.

Thinking levels on the Enterprise Agent Platform page, read the same day, are LOW, MEDIUM, HIGH (the default), and XHIGH. Setting thinking_budget, or thinking_level to MINIMAL, returns HTTP 400. The per-token price does not change with the level. Google has not published how many thinking tokens each level emits.

Where to put the id in a harness, next to Carbon, Argon, and Claude Fable 5.2, is Gemini 4 Hydrogen vs Carbon vs Argon for coding agents. Gemini is the product family. The Carbon id and the HTTP 400 rules that page already documented are Gemini 4 Carbon for agent harnesses.