Gemini 4 Hydrogen
Google's everyday-execution tier: 54 intelligence, above Argon, at 25% lower list price
Benchmark scores
How it is measured →Ranked 7 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 54.0 | 83% | −11.0 pts |
| AA Intelligence Index v4.3.2 | — | — | not indexed |
| GPQA Diamond | — | — | not indexed |
| Terminal-Bench 4.0 Terminal-agent tasks | 53.5 | 80% | −13.7 pts |
| LiveBench Design | — | — | not indexed |
| Vals Index | — | — | not indexed |
Assessment
Launched 11-12 as the third role alongside Carbon and Argon, aimed at everyday execution and high-frequency agent calls: available from day one across all AI Studio tiers, the Gemini API, the Gemini app and AI Mode in Search, with paid Pro/Ultra subscribers also getting it in Antigravity and Sheets. AA's board puts its top xhigh tier at 54, above Argon's high 50 (the board records high 51); $1.5/$7.5 is flat across the 1M-token context with no long-context surcharge, and unlike Carbon and Argon there is no 12-01 step-up price. Two costs are worth seeing: it is extremely verbose, burning 304M output tokens on the full index (median 72.8M), and 13.7 minutes per task is slightly slower than the 13.0-minute median; in exchange one index task costs $3.16, about 60% of Argon's $4.97. Terminal-Bench 4.0 is 53.5%, taken from the same field AA records for every other row (AA's own run), which is a step below Argon's 57.1 — more intelligence, but weaker on terminal-agent work than the tier above it. GPQA is still null in AA's record, so that column stays blank rather than guessed.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| Grok 4.7xAI | 35.5 | $3.00 | 63 | -18.5 |
| Qwen3.8 MaxAlibaba Qwen | 34.0 | $3.00 | 54 | -20.0 |
| GLM-5.3 Open weightsZ.ai | 32.0 | $2.15 | 71 | -22.0 |
| Claude Sonnet 5.5Anthropic | 54.5 | $4.00 | 146 | +0.5 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboardRelated benchmarks
Data sources & methodology
Full methodology & sources →Every number on this page comes from published evaluations. We run no tests of our own and accept no vendor-submitted scores. Main sources:
Evaluation data in the wider industry also comes from public datasets, leaderboard platforms and their evaluation harnesses, private in-house sets, production logs and user feedback, and synthetic data. This site's current rankings use only the parts that are public, independently checkable and clearly defined.