Gemini 4 Hydrogen (xHigh): benchmark results
Google DeepMind model adapted from Gemini 4 Carbon, released publicly through the Gemini API and apps on November 12, 2026. Low, medium, high (default) and xhigh thinking levels; text, image, audio and video inputs. Provider: Google. Released 2026-11-12. Access: API.
Aggregate Rating 1967 ± 14, rank #14 of 2205 rated models, from 76 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Vals AI Tax Agent Bench | 52.31 | Accuracy (%) | 100 |
| AutomationBench-AA - HR - Raw Objective Completion | 82.56 | Raw Objective Completion (%) | 99.6 |
| AutomationBench-AA - Finance - Raw Objective Completion | 95.3 | Raw Objective Completion (%) | 99.2 |
| AA-Omniscience v1.1 Hallucination Rate | 8.48 | Hallucination Rate (%) | 98.6 |
| AutomationBench-AA | 76.14 | Guardrail-Adjusted Objective Completion (%) | 98.5 |
| AutomationBench-AA - Strict | 49.47 | Strict Task Success (%) | 98.5 |
| AA GDP.pdf - Insurance - Criterion Pass Rate | 89.99 | Criterion Pass Rate (%) | 98.1 |
| APEX v1 | 76.9 | Score (%) | 97.4 |
| Vals AI Finance Index | 66.14 | Finance Index (%) | 97.4 |
| AA-Briefcase - Presentation Elo | 1615.29 | Elo | 97 |
| AA GDP.pdf - Legal - Criterion Pass Rate | 86.1 | Criterion Pass Rate (%) | 96.5 |
| AutomationBench-AA - Support - Raw Objective Completion | 92.27 | Raw Objective Completion (%) | 96.2 |
Interactive version: theaggregate.ai/model?slug=gemini-4-hydrogen-xhigh · How It Works · Data refreshed daily, snapshot 2026-11-13.