Atlas

Compare Models

GPT-6.1 Terra vs Muse Wonder

GPT-6.1 Terra leads by 1.7 capability points (≈95% interval −0.7 to +4.1): statistically tied · P(higher) ≈ 92%

Wins 51 of 91 shared benchmarks (39 lost, 1 tied)

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

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

GPT-6.1 TerravsMuse Wondercapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins51 – 39 · 91 shared, 1 tied

Comparing matching setups only (13 benchmarks excluded) · suites counted once
← GPT-6.1 TerraMuse Wonder →

GPT-6.1 TerravsMuse Wonderwins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 23 othersVillage LLM Atlasatlas.village.dev

Margins

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

EBR-bench v4 (card ban, single agent)Games78.131.4−46.7
Terminal-Bench-Science 0.1Science57.625.7−31.9
VoxelBench Text RatingGames23131870−31.6
Visa Agentic Commerce BenchmarkAgents51.182.2+31.1
Furniture Assembly BenchmarkMultimodal79.254.2−25.0
Legal Research BenchLegal26.050.5+24.5
Mystery Game PuzzlesGames74.051.0−23.0
HLAB-Vals-1.1Legal1.723.8+22.1
FrontierMath Tier 4 (v2) - PrivateMath92.770.7−22.0
BioSecBench-Refusal V1 Clean RerunsSafety34.255.3+21.1
DistressBenchSafety65.985.9+20.0
GBENCH One-shot Coding GScoreGames77.658.1−19.5
LMArena Agent Arena Net Improvement (Steering Burden)Agents12.01.8−19.3
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.523.0+18.5
NoCoT-Bench NCRI 15.2General QA13097.7−18.0
Vals Legal IndexIndexes58.875.3+16.5
Tax Agent BenchFinance & Accounting55.071.2+16.1
ChessBench CoherenceGames84.8100.0+15.2
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.151.9+14.8
FrontierCode 2.0 (Artificial Analysis)Code42.327.9−14.4
AA-Briefcase v1.1 EloAgents14011659+14.1
GDPval-AA v2.1Professional Work14761770+13.0
WeirdML v3Science28.115.2−12.9
ChessBench EloGames12511046−12.2
MacroscopeBenchCode76.164.2−11.9
HiL-Bench (Human-in-Loop Benchmark)Agents68.079.7+11.7
LMArena Agent Arena Praise vs ComplaintAgents20.0-1.8−10.9
APEX-1 (Mercor, mean score)Professional Work64.275.1+10.9
Toloka Arena (v3)Agents55.065.7+10.7
FrontierCode 1.1 Main (Artificial Analysis)Code36.046.5+10.4
CritPt v2 (agentic, avg@5)Science81.471.1−10.3
Finance Agent v2.1Finance & Accounting56.065.5+9.6
DAXBench v2Code93.684.0−9.6
FrontierMath Tiers 1-3 (v2) - PrivateMath90.982.1−8.8
ChessBench AccuracyGames78.370.2−8.7
HardGeoBench Avg DistanceMultimodal10911369−8.6
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine30.622.2−8.3
Humanity's Last Exam DiamondGeneral QA47.339.0−8.3
GameBench 2Games85.877.8−8.0
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine16.024.0+8.0
AutomationBench-AA 1.0.6Agents60.268.1+7.9
Vals Finance IndexIndexes53.361.2+7.9
Artificial Analysis Intelligence Index v5.0.1Indexes51.343.6−7.8
Artificial Analysis Intelligence Index v5.0Indexes51.043.3−7.7
HardGeoBench <=100miMultimodal46.338.8−7.5
HardGeoBench <=10miMultimodal35.027.5−7.5
SWE Atlas v1.1 - OverallCode57.764.8+7.1
LMArena Agent Arena Steering BurdenAgents16.62.6−7.0
BuseyBench SVGMultimodal7.36.6−6.8
HardGeoBench Country AccuracyMultimodal58.852.5−6.2
SciCode 1.0.1 (Artificial Analysis)Code55.160.3+5.2
Artificial Analysis Engineering IndexIndexes54.749.6−5.2
DualEntry Accounting AI BenchmarkFinance & Accounting78.283.2+5.0
VIALS-AA (pass^3)Multimodal60.255.3−5.0
Epoch Capabilities Index (ECI)Indexes168162−4.9
Artificial Analysis Strategy & Ops IndexIndexes50.054.5+4.5
FrontierMemory v1.1Agents46.751.1+4.4
Vals Index v2.2Indexes60.564.9+4.3
LMArena Agent Arena Bash ReliabilityAgents13.44.8−4.3
Artificial Analysis Intelligence Index v4.3.2Indexes48.152.4+4.3
Artificial Analysis Healthcare & Medical IndexIndexes44.748.7+4.0
Humanity's Last Exam (Text-Only)General QA46.250.1+3.8
MKA (Artificial Analysis)Professional Work43.847.5+3.7
Terminal-Bench 4.0 (Artificial Analysis)Code41.938.4−3.5
APEX-Accounting 1.1Finance & Accounting63.767.1+3.4
Artificial Analysis Progress Index v1Indexes345336−3.3
VIALS-AA (pass@3)Multimodal71.473.9+2.5
Code Migration v1.1 (Vals Index v2.2 subset)Code72.269.8−2.4
APEX-Agents 1.1Professional Work79.377.0−2.3
AA-Omniscience IndexIndexes29.825.5−2.1
LMArena Agent Arena Confirmed SuccessAgents9.75.4−2.1
Artificial Analysis Finance & Accounting IndexIndexes49.051.1+2.1
TaxCalcBench TY25Finance & Accounting24.022.0−2.0
GDP.pdf (Artificial Analysis, All-pass)Professional Work28.630.6+2.0
Vending-Bench 2Games91079432+2.0
GDPxProfessional Work8.56.8−1.7
LiveBench 2.0Agents79.778.1−1.6
SpatialBench V1 Clean RerunsScience73.371.9−1.5
VIALS-AA (average of 3)Multimodal66.064.8−1.2
Artificial Analysis Economics IndexIndexes48.549.5+1.0
Vibe Code Bench v1.1Code84.083.1−0.9
Excel Modeling Benchmark v1.1Finance & Accounting73.372.5−0.8
LMArena Agent Arena Tool HallucinationAgents1.0-0.5−0.8
OTIS Mock AIME 2024-2025Math100.099.4−0.6
Artificial Analysis Legal IndexIndexes46.446.9+0.5
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting85.785.4−0.3
CritPt (avg@5)Science28.028.3+0.3
MMMU ProMultimodal84.784.9+0.2
NoCoT-Bench NCKI (kspine v3)General QA110109−0.2
Prophet Arena Arena Score (Fixed Context)Forecasting85.285.20.0
Terminal-Bench 4.0Code49.549.50.0

GPT-6.1 TerravsMuse Wondernormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 23 othersVillage LLM Atlasatlas.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…