Atlas

Compare Models

Gemini 4 Hydrogen vs Muse Ember

Gemini 4 Hydrogen leads by 3.5 capability points (≈95% interval +1.0 to +6.1): a clear lead over Muse Ember · P(higher) ≈ 100%

Wins 45 of 53 shared benchmarks (8 lost)

Capability vs. relative cost
Gemini 4 HydrogenMuse EmberClosed weightsOpen weights

Benchmark wins45 – 8 · 53 shared

Comparing matching setups only (1 benchmark excluded) · suites counted once
← Gemini 4 HydrogenMuse Ember →
Data: Artificial Analysis, Vals AI, Arena and 7 othersatlas.village.dev

Margins

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

TaxCalcBench TY25Professional › Finance, accounting & tax44.02.0−42.0
Code Migration v1.1 (Vals Index v2.2 subset)Software › Software engineering82.747.1−35.6
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge58.825.1−33.7
Terminal-Bench-Science 0.1Science › Multidisciplinary49.517.6−31.9
GDP.xlsxProfessional › General knowledge work40.915.4−25.5
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents13.51.0−24.5
Ramp SWE-Bench v2Software › Software engineering71.651.8−19.8
AA-Omniscience v1.1 IndexKnowledge & reasoning › General knowledge46.37.7−19.3
SimpleBenchKnowledge & reasoning › General knowledge91.772.7−19.0
HiL-Bench (Human-in-Loop Benchmark)Tool & computer use › General agents55.374.0+18.7
CritPt v2 (agentic, avg@5)Science › Physical sciences80.662.0−18.6
Artificial Analysis Engineering IndexScience › Engineering60.141.8−18.3
Artificial Analysis Legal IndexProfessional › Legal56.738.5−18.3
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering53.535.4−18.2
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering39.022.0−17.0
DrugDiscoveryBenchScience › Life sciences69.152.4−16.7
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents29.3-2.7−16.0
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following73.758.8−14.9
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations44.629.8−14.8
Terminal-Bench 4.0Software › Software engineering54.039.4−14.6
BuseyBench SVGPerception & media › Generation6.95.4−14.6
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax38.524.2−14.3
Artificial Analysis Economics IndexComposite indexes55.041.0−14.1
HLAB-Vals-1.1Professional › Legal11.325.0+13.8
Artificial Analysis Finance & Accounting IndexProfessional › Finance, accounting & tax58.144.6−13.5
Artificial Analysis Coding Agent Index v2.1Software › Software engineering52.339.7−12.6
LMArena Code Arena WebDevSoftware › Frontend & design17031608−12.6
Artificial Analysis Healthcare & Medical IndexHealth & medicine53.341.3−12.1
GBENCH One-shot Coding GScoreGames66.856.4−10.4
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work32.822.8−10.0
Artificial Analysis Strategy & Ops IndexProfessional › General knowledge work57.748.7−9.1
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine27.218.3−8.9
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents15.7-2.0−8.9
Vals Index v2.2Composite indexes68.560.5−8.0
Chartography (November 2026 rescoring)Perception & media › Documents & charts87.279.3−7.9
Tax Agent BenchProfessional › Finance, accounting & tax78.471.4−7.1
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine24.017.2−6.8
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations74.468.1−6.4
LMArena Agent Arena Steering BurdenTool & computer use › General agents17.44.8−6.3
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents-4.56.6+5.6
FORTRESS Average Risk Score (ARS)Safety & alignment17.511.9+5.5
Vals Finance IndexProfessional › Finance, accounting & tax66.160.9−5.2
Artificial Analysis Intelligence Index v5.0.1Composite indexes41.236.0−5.2
GDP.pdf (November 2026 rescoring)Professional › General knowledge work54.249.2−5.0
Legal Research BenchProfessional › Legal53.849.0−4.8
Vibe Code Bench v1.1Software › Frontend & design87.883.7−4.1
Professional Reasoning Benchmark - FinanceProfessional › Finance, accounting & tax52.356.0+3.7
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax75.872.2−3.6
Professional Reasoning Benchmark - LegalProfessional › Legal53.155.6+2.5
AA-Briefcase v1.1 EloProfessional › General knowledge work15691526−2.3
Vals Legal IndexProfessional › Legal73.675.7+2.1
Finance Agent v2.1Professional › Finance, accounting & tax65.163.8−1.3
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling0.70.9+0.1
Data: Artificial Analysis, Vals AI, Arena and 7 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…