Atlas

Compare Models

GPT-6.1 Terra vs Muse Ember

GPT-6.1 Terra leads by 3.1 capability points (≈95% interval +0.6 to +5.5): a clear lead over Muse Ember · P(higher) ≈ 99%

Wins 36 of 51 shared benchmarks (15 lost)

Capability vs. relative cost
GPT-6.1 TerraMuse EmberClosed weightsOpen weights

Benchmark wins36 – 15 · 51 shared

Comparing matching setups only (9 benchmarks excluded) · suites counted once
← GPT-6.1 TerraMuse Ember →
Data: Artificial Analysis, Vals AI, Arena and 10 othersatlas.village.dev

Margins

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

Terminal-Bench-Science 0.1Science › Multidisciplinary57.620.0−37.6
DAXBench v2Software › Data & ML93.662.8−30.9
Code Migration v1.1 (Vals Index v2.2 subset)Software › Software engineering72.247.1−25.1
HLAB-Vals-1.1Professional › Legal1.725.0+23.3
Legal Research BenchProfessional › Legal26.049.0+23.1
KradleBenchGames48.625.7−22.9
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge47.325.1−22.2
TaxCalcBench TY25Professional › Finance, accounting & tax24.02.0−22.0
GBENCH One-shot Coding GScoreGames76.456.4−20.0
CritPt v2 (agentic, avg@5)Science › Physical sciences81.462.3−19.1
BuseyBench SVGPerception & media › Generation7.35.4−18.8
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents10.51.0−18.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Professional › Legal4.522.7+18.2
Vals Legal IndexProfessional › Legal58.875.7+16.9
Tax Agent BenchProfessional › Finance, accounting & tax55.071.4+16.3
τ³-Banking (Mercor run, bm25_grep retrieval)Tool & computer use › Customer service & operations37.153.3+16.2
Artificial Analysis Intelligence Index v5.0.1Composite indexes51.336.0−15.3
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering42.327.2−15.2
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering55.640.9−14.6
Artificial Analysis Engineering IndexScience › Engineering54.741.8−13.0
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine30.618.3−12.2
AA-Omniscience v1.1 IndexKnowledge & reasoning › General knowledge29.67.7−10.9
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents17.9-2.7−10.3
Terminal-Bench 4.0Software › Software engineering49.539.4−10.1
Artificial Analysis Legal IndexProfessional › Legal46.438.5−7.9
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations60.268.1+7.9
Finance Agent v2.1Professional › Finance, accounting & tax56.063.8+7.9
Vals Finance IndexProfessional › Finance, accounting & tax53.360.9+7.6
Artificial Analysis Economics IndexComposite indexes48.541.0−7.5
AA-Briefcase v1.1 EloProfessional › General knowledge work14001520+6.6
FORTRESS Average Risk Score (ARS)Safety & alignment18.211.9+6.3
APEX-1 (Mercor, mean score)Professional › General knowledge work64.270.4+6.2
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering60.654.4−6.2
HiL-Bench (Human-in-Loop Benchmark)Tool & computer use › General agents68.074.0+6.0
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work28.622.8−5.8
Artificial Analysis Progress Index v1Composite indexes345329−5.8
DrugDiscoveryBenchScience › Life sciences58.152.4−5.7
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents8.5-2.0−5.3
Artificial Analysis Finance & Accounting IndexProfessional › Finance, accounting & tax49.044.6−4.4
LMArena Agent Arena Steering BurdenTool & computer use › General agents13.24.8−4.2
APEX-Agents 1.1Professional › General knowledge work79.375.7−3.6
Artificial Analysis Healthcare & Medical IndexHealth & medicine44.741.3−3.4
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents12.76.6−3.0
SWE Atlas v1.1 - OverallSoftware › Software engineering57.756.0−1.7
Artificial Analysis Strategy & Ops IndexProfessional › General knowledge work50.048.7−1.3
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax73.372.2−1.2
SimpleBenchKnowledge & reasoning › General knowledge72.272.7+0.5
Vibe Code Bench v1.1Software › Frontend & design84.083.7−0.3
ParseBench-2Perception & media › Documents & charts74.374.5+0.3
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.20.9−0.1
Vals Index v2.2Composite indexes60.560.5−0.1
Data: Artificial Analysis, Vals AI, Arena and 10 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…