Atlas

Compare Models

GPT-6.1 Terra vs Grok 4.6

GPT-6.1 Terra leads by 5.3 capability points (≈95% interval +3.0 to +7.6): a clear lead over Grok 4.6 · P(higher) ≈ 100%

Wins 63 of 88 shared benchmarks (23 lost, 2 tied)

Capability vs. relative cost
GPT-6.1 TerraGrok 4.6Closed 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 TerravsGrok 4.6capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins63 – 23 · 88 shared, 2 tied

Comparing matching setups only (9 benchmarks excluded) · suites counted once
← GPT-6.1 TerraGrok 4.6 →

GPT-6.1 TerravsGrok 4.6wins by categoryData Nov 15, 2026Data: benchmarks.bio, Artificial Analysis, Epoch AI and 24 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.6 minus GPT-6.1 Terra.

FrontierMath Tier 4 (v2) - PrivateMath92.731.7−61.0
ChartographyMultimodal63.220.5−42.7
Mystery Game PuzzlesGames74.034.0−40.0
Furniture Assembly BenchmarkMultimodal79.240.0−39.2
Terminal-Bench-Science 0.1Science41.92.9−39.0
LLM Chess EloGames1391443−38.9
ALE-BenchCode28751822−37.3
Terminal-Bench 4.0 (Artificial Analysis)Code46.517.2−29.3
BuseyBench SVGMultimodal7.34.6−26.7
Humanity's Last Exam DiamondGeneral QA47.321.5−25.8
Terminal-Bench 4.0Code50.325.3−25.0
FrontierMath Tiers 1-3 (v2) - PrivateMath90.966.0−24.9
NoCoT-Bench NCRI 15.2General QA13086.8−24.0
TissueMicroscopyBench V1 (49 evaluations)Science25.92.7−23.1
LMArena Agent Arena Net Improvement (Steering Burden)Agents12.0-0.2−23.0
GBENCH One-shot Coding GScoreGames77.655.4−22.2
Legal Research BenchLegal26.048.1+22.1
HardGeoBench Avg DistanceMultimodal10911807−22.1
ChessBench EloGames1251891−21.4
RareDiseaseBench V1 (35 evaluations)Science25.74.8−21.0
ChessBenchGames64.043.3−20.7
Artificial Analysis Intelligence Index v5.0.1Indexes51.331.0−20.3
VoxelBench Text RatingGames23132050−18.7
GameBench 2Games85.868.7−17.1
FrontierCode 2.0 (Artificial Analysis)Code36.919.8−17.1
Tax Agent BenchFinance & Accounting55.071.4+16.3
TxBench Oligo Discovery V1 (113 evaluations)Science30.446.0+15.6
ChessBench CoherenceGames84.869.2−15.6
HardGeoBench <=10miMultimodal35.020.0−15.0
HardGeoBench Country AccuracyMultimodal58.843.8−15.0
CritPt v2 (agentic, avg@5)Science77.162.3−14.9
SpatialBench-Long V1 Clean RerunsScience40.526.2−14.3
HardGeoBench <=100miMultimodal46.332.5−13.7
BioSecBench-Refusal V1 Clean RerunsSafety34.246.9+12.7
TxBench Antibody Discovery V1 (100 evaluations)Science40.753.0+12.3
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine30.618.3−12.2
AA-Briefcase v1.1 EloAgents13341540+11.3
LMArena Agent Arena Praise vs ComplaintAgents20.0-2.5−11.3
ChessBench AccuracyGames78.367.8−11.3
HealthBench Professional (Length-Adjusted)Health & Medicine62.751.7−11.0
GDP.pdf (Artificial Analysis, All-pass)Professional Work28.617.8−10.8
GDPval-AA v2.1Professional Work14161658+10.7
BioSecBench-Structure V1 (58 evaluations)Science25.915.5−10.3
LMArena Agent Arena Steering BurdenAgents16.6-1.8−9.2
Blueprint-Bench 2Multimodal0.40.3−8.9
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.145.7+8.6
Artificial Analysis Progress Index v1Indexes345322−8.3
CritPt (avg@5)Science28.019.7−8.3
Epoch Capabilities Index (ECI)Indexes168158−8.2
SimpleQA Verified (Epoch anti-abstention)General QA57.349.3−8.0
Vibe Code Bench v1.1Code84.076.2−7.8
BioSecBench-Surveillance V1 Clean RerunsScience39.947.5+7.6
DAXBench v2Code93.686.2−7.4
MacroscopeBenchCode82.375.1−7.2
AutomationBench-AA 1.0.6Agents60.267.0+6.8
LMArena Agent Arena Confirmed SuccessAgents9.7-2.6−6.1
SpatialBench V1 Clean RerunsScience73.367.3−6.1
TxBench Preclinical Pharmacology V1 Clean RerunsScience55.761.7+6.0
VariantBench V1 Clean RerunsScience34.840.4+5.6
BioSecBench-Function V1 Clean RerunsScience39.044.0+5.0
LiveBench 2.0Agents79.774.9−4.7
MMMU ProMultimodal84.780.5−4.2
Artificial Analysis Intelligence Index v4.3.2Indexes48.144.3−3.8
APEX-1 (Mercor, mean score)Professional Work64.267.7+3.5
APEX-Accounting 1.1Finance & Accounting63.760.3−3.4
Humanity's Last Exam (Text-Only)General QA46.242.9−3.3
TxBench-Antibody-Development V1 (60 evaluations)Science22.826.1+3.3
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine16.018.7+2.7
ScBench-Long V1 Clean RerunsScience18.816.5−2.3
APEX-Agents 1.1Professional Work79.377.3−2.0
Design Arena Overall Code CategoriesCode13001288−2.0
NeuroscienceBench V1 (91 evaluations)Science22.320.5−1.8
LMArena Agent Arena Bash ReliabilityAgents13.410.0−1.7
AA-LCR v1.1Search82.781.0−1.7
NoCoT-Bench NCKI (kspine v3)General QA110112+1.4
SciCode 1.0.1 (Artificial Analysis)Code55.156.5+1.4
Toloka Arena (v3)Agents55.056.0+1.0
ScBench V1 Clean RerunsScience58.359.1+0.9
OTIS Mock AIME 2024-2025Math100.099.2−0.8
Vending-Bench 2Games91079047−0.4
AA-Omniscience IndexIndexes29.830.5+0.4
GPQA DiamondGeneral QA94.394.0−0.3
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science11.211.5+0.3
EpiBench V1 Clean RerunsScience23.323.6+0.3
Prophet Arena Arena Score (Fixed Context)Forecasting85.285.0−0.2
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting85.785.6−0.1
LMArena Agent Arena Tool HallucinationAgents1.01.00.0
MetagenomicsBench V1Science38.738.70.0

GPT-6.1 TerravsGrok 4.6normalised score marginsData Nov 15, 2026Data: benchmarks.bio, Artificial Analysis, Epoch AI and 24 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…