Gemini 4 Hydrogen vs Grok 4.6
Gemini 4 Hydrogen leads by 6.0 capability points (≈95% interval +3.7 to +8.3): a clear lead over Grok 4.6 · P(higher) ≈ 100%
Wins 48 of 52 shared benchmarks (4 lost)
Capability vs. relative cost
Gemini 4 HydrogenGrok 4.6Closed weightsOpen weights
Benchmark wins48 – 4 · 52 shared
Comparing matching setups only · suites counted once
Area← Gemini 4 HydrogenGrok 4.6 →
Data: Artificial Analysis, Epoch AI, Arena and 15 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.6 minus Gemini 4 Hydrogen.
| FrontierMath Tier 4 (v2) - Private | Math | 87.8 | 31.7 | −56.1 |
| Gray Swan IPI Attack Success (k=15) | Safety & alignment | 0.0 | 51.8 | −51.8 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 49.5 | 2.9 | −46.7 |
| Chartography | Perception & media › Documents & charts | 66.6 | 20.5 | −46.1 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 58.8 | 21.5 | −37.3 |
| Terminal-Bench 4.0 | Software › Software engineering | 54.0 | 17.2 | −36.9 |
| TxBench-Antibody-Development V1 (60 evaluations) | Science › Life sciences | 60.6 | 26.1 | −34.4 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 54.0 | 21.2 | −32.8 |
| Mystery Game Puzzles | Games | 65.0 | 34.0 | −31.0 |
| FrontierSWE V2 | Software › Software engineering | 53.1 | 25.3 | −27.8 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 13.5 | -0.6 | −27.5 |
| Chess Puzzles (task version 2.0) | Games | 60.0 | 33.0 | −27.0 |
| ARC-AGI-2 | Knowledge & reasoning › Abstract reasoning | 92.5 | 67.1 | −25.4 |
| BuseyBench SVG | Perception & media › Generation | 6.9 | 4.6 | −22.5 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 88.1 | 66.0 | −22.1 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 70.9 | 49.3 | −21.6 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 39.0 | 19.8 | −19.2 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 80.6 | 62.3 | −18.3 |
| SWE-Marathon v1.1 | Software › Software engineering | 49.4 | 31.9 | −17.5 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 38.5 | 21.0 | −17.5 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 57.5 | 40.0 | −17.5 |
| CursorBench 4.1 | Software › Software engineering | 61.4 | 44.3 | −17.1 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 29.3 | -3.5 | −16.4 |
| SimpleBench | Knowledge & reasoning › General knowledge | 91.7 | 75.9 | −15.8 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 32.8 | 17.8 | −15.0 |
| GBENCH One-shot Coding GScore | Games | 66.8 | 52.3 | −14.5 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 77.0 | 63.5 | −13.5 |
| CursorBench 4.0 | Software › Software engineering | 54.3 | 41.4 | −12.9 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 87.8 | 76.2 | −11.5 |
| LMArena Code Arena WebDev | Software › Frontend & design | 1703 | 1618 | −11.3 |
| ARC-AGI-1 Semi-Private | Knowledge & reasoning › Abstract reasoning | 98.5 | 87.5 | −11.0 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 46.3 | 24.9 | −10.7 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 41.2 | 31.0 | −10.2 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 17.4 | -2.5 | −10.0 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 24.0 | 14.8 | −9.2 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 15.7 | -2.5 | −9.1 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 27.2 | 18.3 | −8.9 |
| TxBench Oligo Discovery V1 (113 evaluations) | Science › Life sciences | 54.3 | 46.0 | −8.3 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 74.4 | 67.0 | −7.4 |
| Epoch Capabilities Index (ECI) | Composite indexes | 167 | 158 | −7.2 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 78.4 | 71.4 | −7.1 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | -4.5 | 9.1 | +6.8 |
| Legal Research Bench | Professional › Legal | 53.8 | 48.1 | −5.8 |
| LMArena Text Arena Elo | Language & writing › General assistance | 1490 | 1457 | −5.8 |
| TxBench Antibody Discovery V1 (100 evaluations) | Science › Life sciences | 57.7 | 53.0 | −4.7 |
| Bug Hunt Bench | Software › Software engineering | 31.1 | 27.3 | −3.8 |
| GPQA Diamond | Science › Multidisciplinary | 96.8 | 94.0 | −2.8 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1569 | 1540 | −1.6 |
| micro1 PharmaBench | Professional › General knowledge work | 11.4 | 12.5 | +1.1 |
| OTIS Mock AIME 2024-2025 | Math | 100.0 | 99.2 | −0.8 |
| TxBench Preclinical Pharmacology V1 Clean Reruns | Science › Life sciences | 61.3 | 61.7 | +0.3 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 0.7 | 1.2 | +0.2 |
Data: Artificial Analysis, Epoch AI, Arena and 15 othersatlas.village.dev