Atlas

Compare Models

GPT-6 Astra vs Muse Ember

GPT-6 Astra leads by 4.9 capability points (≈95% interval +2.4 to +7.5): a clear lead over Muse Ember · P(higher) ≈ 100%

Wins 40 of 57 shared benchmarks (17 lost)

Capability vs. relative cost
GPT-6 AstraMuse EmberClosed weightsOpen weights

Benchmark wins40 – 17 · 57 shared

Comparing matching setups only (9 benchmarks excluded) · suites counted once
← GPT-6 AstraMuse Ember →
Data: Artificial Analysis, Surge AI, Vals AI and 12 othersatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Ember minus GPT-6 Astra.

TaxCalcBench TY25Professional › Finance, accounting & tax46.02.0−44.0
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge63.825.1−38.7
Terminal-Bench-Science 0.1Science › Multidisciplinary54.817.6−37.1
DAXBench v2Software › Data & ML96.862.8−34.0
Code Migration v1.1 (Vals Index v2.2 subset)Software › Software engineering78.047.1−30.9
KradleBenchGames56.125.7−30.4
FoodTruck BenchBusiness simulations5776134639−29.7
RiemannBench (November 2026 rescoring)Math81.653.6−28.0
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents14.61.0−26.5
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering59.635.4−24.2
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine41.718.3−23.3
GBENCH One-shot Coding GScoreGames78.456.4−22.0
CritPt v2 (agentic, avg@5)Science › Physical sciences83.162.3−20.9
Artificial Analysis Intelligence Index v5.0.1Composite indexes56.936.0−20.8
Terminal-Bench 4.0Software › Software engineering59.639.4−20.2
EnterpriseBench: CoreCraftTool & computer use › Customer service & operations42.662.1+19.5
BuseyBench SVGPerception & media › Generation7.35.4−19.2
AA-Omniscience v1.1 IndexKnowledge & reasoning › General knowledge44.67.7−18.4
HLAB-Vals-1.1Professional › Legal7.125.0+17.9
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents32.3-2.7−17.5
DrugDiscoveryBenchScience › Life sciences68.752.4−16.3
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering42.427.2−15.2
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Professional › Legal9.222.7+13.5
SWE Atlas v1.1 - OverallSoftware › Software engineering69.456.0−13.4
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax37.324.2−13.1
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following70.858.8−12.0
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine28.417.2−11.2
SimpleBenchKnowledge & reasoning › General knowledge83.672.7−10.9
Legal Research BenchProfessional › Legal39.449.0+9.6
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work32.222.8−9.4
Chartography (November 2026 rescoring)Perception & media › Documents & charts88.679.3−9.3
HiL-Bench (Human-in-Loop Benchmark)Tool & computer use › General agents65.374.0+8.7
Professional Reasoning Benchmark - FinanceProfessional › Finance, accounting & tax47.556.0+8.4
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents14.9-2.0−8.4
Tax Agent BenchProfessional › Finance, accounting & tax63.371.4+8.0
GDP.xlsxProfessional › General knowledge work22.915.4−7.5
Artificial Analysis Progress Index v1Composite indexes350329−7.3
Vals Legal IndexProfessional › Legal68.475.7+7.3
Professional Reasoning Benchmark - LegalProfessional › Legal48.455.6+7.2
τ³-Banking (Mercor run, bm25_grep retrieval)Tool & computer use › Customer service & operations46.753.3+6.6
OSWorld 2.1 full - reported partial reward, step limit unspecifiedTool & computer use › Computer & browser use75.869.5−6.3
Vals Finance IndexProfessional › Finance, accounting & tax55.660.9+5.3
LMArena Agent Arena Steering BurdenTool & computer use › General agents13.94.8−4.6
APEX-SWESoftware › Software engineering48.652.9+4.3
Finance Agent v2.1Professional › Finance, accounting & tax60.463.8+3.4
AA-Briefcase v1.1 EloProfessional › General knowledge work15691520−2.7
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations31.429.8−1.6
GDP.pdf (November 2026 rescoring)Professional › General knowledge work50.449.2−1.2
APEX-1 (Mercor, mean score)Professional › General knowledge work69.370.4+1.1
APEX-Agents 1.1Professional › General knowledge work75.175.7+0.6
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax72.772.2−0.5
OSWorld 2.1 full - reported binary completion, step limit unspecifiedTool & computer use › Computer & browser use34.333.8−0.5
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations68.568.1−0.4
FORTRESS Average Risk Score (ARS)Safety & alignment12.111.9+0.2
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering54.654.4−0.2
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.20.9−0.1
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents6.56.6+0.1
Data: Artificial Analysis, Surge AI, Vals 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…