Atlas

Compare Models

Gemini 4 Hydrogen vs GPT-6.1 Terra

Gemini 4 Hydrogen leads by 0.5 capability points (≈95% interval −2.1 to +3.0): statistically tied · P(higher) ≈ 64%

Wins 45 of 61 shared benchmarks (15 lost, 1 tied)

Capability vs. relative cost
Gemini 4 HydrogenGPT-6.1 TerraClosed weightsOpen weights

Benchmark wins45 – 15 · 61 shared, 1 tied

Comparing matching setups only (31 benchmarks excluded) · suites counted once
← Gemini 4 HydrogenGPT-6.1 Terra →
Data: Artificial Analysis, Vals AI, Epoch AI and 12 othersatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Terra minus Gemini 4 Hydrogen.

TxBench-Antibody-Development V1 (60 evaluations)Science › Life sciences60.622.8−37.8
Legal Research BenchProfessional › Legal53.826.0−27.9
TxBench Oligo Discovery V1 (113 evaluations)Science › Life sciences54.330.4−23.9
Tax Agent BenchProfessional › Finance, accounting & tax78.455.0−23.4
Furniture Assembly BenchmarkPerception & media › Spatial & visual reasoning57.579.2+21.7
TaxCalcBench TY25Professional › Finance, accounting & tax44.024.0−20.0
SimpleBenchKnowledge & reasoning › General knowledge91.772.2−19.5
TxBench Antibody Discovery V1 (100 evaluations)Science › Life sciences57.740.7−17.0
Vals Legal IndexProfessional › Legal73.658.8−14.9
SpireClimbGames7.722.3+14.7
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations74.460.2−14.3
SimpleQA Verified (Epoch anti-abstention)Knowledge & reasoning › General knowledge70.957.3−13.6
AA-Briefcase v1.1 EloProfessional › General knowledge work15691334−12.9
Vals Finance IndexProfessional › Finance, accounting & tax66.153.3−12.8
HiL-Bench (Human-in-Loop Benchmark)Tool & computer use › General agents55.368.0+12.7
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge58.847.3−11.5
DrugDiscoveryBenchScience › Life sciences69.158.1−11.0
Code Migration v1.1 (Vals Index v2.2 subset)Software › Software engineering82.772.2−10.5
Artificial Analysis Legal IndexProfessional › Legal56.746.4−10.4
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering70.860.6−10.2
Artificial Analysis Intelligence Index v5.0.1Composite indexes41.251.3+10.1
GBENCH One-shot Coding GScoreGames66.876.4+9.6
HLAB-Vals-1.1Professional › Legal11.31.7−9.6
Finance Agent v2.1Professional › Finance, accounting & tax65.156.0−9.2
Artificial Analysis Finance & Accounting IndexProfessional › Finance, accounting & tax58.149.0−9.1
Mystery Game PuzzlesGames65.074.0+9.0
Artificial Analysis Healthcare & Medical IndexHealth & medicine53.344.7−8.6
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents-4.512.7+8.6
AA-Omniscience v1.1 IndexKnowledge & reasoning › General knowledge46.329.6−8.3
Vals Index v2.2Composite indexes68.560.5−8.0
Artificial Analysis Strategy & Ops IndexProfessional › General knowledge work57.750.0−7.7
EBR-bench v4 (card ban, single agent)Games70.578.1+7.6
Terminal-Bench-Science 0.1Science › Multidisciplinary49.541.9−7.6
Gray Swan IPI Attack Success (k=15)Safety & alignment0.07.3−7.3
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering53.546.5−7.1
Artificial Analysis Economics IndexComposite indexes55.048.5−6.6
MultiChallengeLanguage & writing › Conversation80.674.2−6.4
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents13.510.5−5.9
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents29.317.9−5.7
TxBench Preclinical Pharmacology V1 Clean RerunsScience › Life sciences61.355.7−5.7
Artificial Analysis Engineering IndexScience › Engineering60.154.7−5.4
DualEntry Accounting AI BenchmarkProfessional › Finance, accounting & tax83.278.2−5.0
FrontierMath Tier 4 (v2) - PrivateMath87.892.7+4.9
Terminal-Bench 4.0Software › Software engineering54.049.5−4.5
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work32.828.6−4.2
BuseyBench SVGPerception & media › Generation6.97.3+4.2
Vibe Code Bench v1.1Software › Frontend & design87.884.0−3.8
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents15.78.5−3.6
CritPt v2 (agentic, avg@5)Science › Physical sciences80.677.1−3.4
ChartographyPerception & media › Documents & charts66.663.2−3.4
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine27.230.6+3.3
FrontierMath Tiers 1-3 (v2) - PrivateMath88.190.9+2.8
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax75.873.3−2.5
GPQA DiamondScience › Multidisciplinary96.894.3−2.5
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering39.036.9−2.1
LMArena Agent Arena Steering BurdenTool & computer use › General agents17.413.2−2.1
LiveBench 2.0 (2026-11-06 release)Composite indexes77.078.6+1.6
Epoch Capabilities Index (ECI)Composite indexes167168+0.7
FORTRESS Average Risk Score (ARS)Safety & alignment17.518.2−0.7
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling0.71.2+0.2
OTIS Mock AIME 2024-2025Math100.0100.00.0
Data: Artificial Analysis, Vals AI, Epoch AI and 12 othersatlas.village.dev
Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…