Atlas

Compare Models

GPT-5.5 vs GPT-6.1 Terra

GPT-6.1 Terra leads by 5.3 capability points (≈95% interval +3.0 to +7.5): a clear lead over GPT-5.5 · P(higher) ≈ 100%

Wins 48 of 70 shared benchmarks (20 lost, 2 tied)

Capability vs. relative cost
GPT-5.5GPT-6.1 TerraClosed 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-5.5vsGPT-6.1 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins20 – 48 · 70 shared, 2 tied

Comparing matching setups only (4 benchmarks excluded) · suites counted once
← GPT-5.5GPT-6.1 Terra →

GPT-5.5vsGPT-6.1 Terrawins by categoryData Nov 15, 2026Data: Artificial Analysis, benchmarks.bio, Epoch AI and 22 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Terra minus GPT-5.5.

ExploitBench - release-reported exploit progressCybersecurity47.997.1+49.2
Furniture Assembly BenchmarkMultimodal44.279.2+35.0
ChartographyMultimodal31.063.2+32.2
Terminal-Bench 4.0 (Artificial Analysis)Code14.646.5+31.8
HiL-Bench (Human-in-Loop Benchmark)Agents39.768.0+28.3
VoxelBench Text RatingGames19252313+27.6
ALE-BenchCode21082875+27.2
GameBench 2Games59.985.8+25.9
LMArena Code Arena WebDevCode15121679+21.8
FrontierMath Tier 4 (v2) - PrivateMath72.592.7+20.2
Humanity's Last Exam DiamondGeneral QA27.947.3+19.4
Mystery Game PuzzlesGames56.074.0+18.0
NoCoT-Bench NCRI 15.2General QA100130+16.6
GBENCH One-shot Coding GScoreGames62.277.6+15.4
ChessBench CoherenceGames100.084.8−15.2
Legal Research BenchLegal40.426.0−14.4
Vibe Code Bench v1.1Code69.884.0+14.2
Agents' Last Exam Overall ScoreAgents46.960.0+13.1
AutomationBench-AA 1.0.6Agents47.360.2+12.9
BuseyBench SVGMultimodal6.07.3+12.8
LiveBench 2.0Agents68.779.7+11.0
AA-Briefcase v1.1 EloAgents11351334+10.9
BioSecBench-Function V1 Clean RerunsScience28.339.0+10.7
Artificial Analysis Intelligence Index v4.3.2Indexes38.448.1+9.8
Vending-Bench 2Games75249107+9.6
SpatialBench V1 Clean RerunsScience64.173.3+9.3
APEX-Agents 1.1Professional Work70.379.3+9.0
Artificial Analysis Progress Index v1Indexes320345+8.9
SpatialBench-Long V1 Clean RerunsScience32.140.5+8.3
HealthBench Hard (updated implementation)Health & Medicine33.826.0−7.8
TxBench Antibody Discovery V1 (100 evaluations)Science33.040.7+7.7
APEX-Accounting 1.1Finance & Accounting56.363.7+7.4
HealthBench (updated implementation)Health & Medicine58.451.1−7.3
BioSecBench-Surveillance V1 Clean RerunsScience47.139.9−7.2
Design Arena Overall Code CategoriesCode12571300+7.1
HealthBench ProfessionalHealth & Medicine57.264.2+7.0
Epoch Capabilities Index (ECI)Indexes160168+6.5
GDP.pdf (Artificial Analysis, All-pass)Professional Work22.628.6+6.0
Blueprint-Bench 2Multimodal0.40.4+5.9
ChessBench EloGames13491251−5.8
LLM Chess EloGames15321391−5.8
SimpleQA Verified (Epoch anti-abstention)General QA63.057.3−5.7
FrontierMath Tiers 1-3 (v2) - PrivateMath85.390.9+5.6
Tax Agent BenchFinance & Accounting60.555.0−5.4
DAXBench v2Code88.393.6+5.3
VulcanBench Frontier v4Code78.583.7+5.2
EpiBench V1 Clean RerunsScience28.323.3−5.0
AA-Omniscience IndexIndexes20.529.8+4.6
VariantBench V1 Clean RerunsScience30.534.8+4.2
APEX-1 (Mercor, mean score)Professional Work68.464.2−4.2
GPQA DiamondGeneral QA90.794.3+3.7
TxBench Oligo Discovery V1 (113 evaluations)Science33.930.4−3.5
MMMU ProMultimodal81.284.7+3.5
ScBench-Long V1 Clean RerunsScience21.618.8−2.8
GDPval-AA v2.1Professional Work13571416+2.6
CritPt (avg@5)Science25.428.0+2.6
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine28.930.6+1.7
AA-LCR v1.1Search84.382.7−1.7
HealthBench (length-adjusted)Health & Medicine56.555.1−1.4
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine14.716.0+1.3
Humanity's Last Exam (Text-Only)General QA45.046.2+1.2
ScBench V1 Clean RerunsScience57.158.3+1.2
SciCode 1.0.1 (Artificial Analysis)Code56.155.1−1.0
HealthBench Consensus (updated implementation)Health & Medicine95.794.8−0.9
NoCoT-Bench NCKI (kspine v3)General QA111110−0.7
ChessBench AccuracyGames78.878.3−0.5
TxBench Preclinical Pharmacology V1 Clean RerunsScience55.355.7+0.3
HealthBench Hard (length-adjusted)Health & Medicine31.531.4−0.1
OTIS Mock AIME 2024-2025Math100.0100.00.0
TaxCalcBench TY25Finance & Accounting24.024.00.0

GPT-5.5vsGPT-6.1 Terranormalised score marginsData Nov 15, 2026Data: Artificial Analysis, benchmarks.bio, Epoch AI and 22 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…