Atlas

Compare Models

Gemini 4 Carbon vs GPT-6.1 Terra

Gemini 4 Carbon leads by 2.6 capability points (≈95% interval −0.1 to +5.4): statistically tied · P(higher) ≈ 97%

Wins 49 of 55 shared benchmarks (5 lost, 1 tied)

Capability vs. relative cost
Gemini 4 CarbonGPT-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.

Gemini 4 CarbonvsGPT-6.1 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins49 – 5 · 55 shared, 1 tied

Comparing matching setups only (9 benchmarks excluded) · suites counted once
← Gemini 4 CarbonGPT-6.1 Terra →

Gemini 4 CarbonvsGPT-6.1 Terrawins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Arena and 8 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 Gemini 4 Carbon.

TxBench-Antibody-Development V1 (60 evaluations)Science61.122.8−38.3
Terminal-Bench-Science 0.1Science62.424.8−37.6
Legal Research BenchLegal55.826.0−29.8
Terminal-Bench 4.0 (Artificial Analysis)Code60.632.3−28.3
TxBench Oligo Discovery V1 (113 evaluations)Science55.830.4−25.4
TxBench Antibody Discovery V1 (100 evaluations)Science65.040.7−24.3
Humanity's Last Exam DiamondGeneral QA69.147.3−21.8
Tax Agent BenchFinance & Accounting76.455.0−21.4
AA-Briefcase v1.1 EloAgents16111265−18.9
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.737.1−18.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal23.04.5−18.5
AutomationBench-AA 1.0.6Agents77.860.2−17.6
Vals Legal IndexIndexes75.958.8−17.2
GDPval-AA v2.1Professional Work17331353−16.7
Blueprint-Bench 2Multimodal0.60.4−16.5
HLAB-Vals-1.1Legal17.51.7−15.8
Artificial Analysis Legal IndexIndexes62.046.4−15.6
Vals Finance IndexIndexes67.053.3−13.7
Terminal-Bench 4.0Code63.149.5−13.6
Artificial Analysis Finance & Accounting IndexIndexes62.649.0−13.6
Artificial Analysis Healthcare & Medical IndexIndexes56.844.7−12.1
AA-Omniscience v1.1 IndexIndexes53.529.6−12.0
Finance Agent v2.1Finance & Accounting67.256.0−11.2
Artificial Analysis Economics IndexIndexes59.648.5−11.1
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine19.430.6+11.1
APEX-Agents 1.1Professional Work90.479.3−11.1
Artificial Analysis Strategy & Ops IndexIndexes60.950.0−10.9
MKA (Artificial Analysis)Professional Work54.343.8−10.6
Artificial Analysis Engineering IndexIndexes65.354.7−10.6
LMArena Agent Arena Praise vs ComplaintAgents41.020.0−10.5
TxBench Preclinical Pharmacology V1 Clean RerunsScience66.055.7−10.3
LMArena Agent Arena Net Improvement (Steering Burden)Agents17.012.0−9.4
LMArena Agent Arena Bash ReliabilityAgents-5.213.4+9.3
ChartographyMultimodal72.563.2−9.3
Artificial Analysis Intelligence Index v5.0.1Indexes60.051.3−8.7
Code Migration v1.1 (Vals Index v2.2 subset)Code80.172.2−7.9
APEX-Accounting 1.1Finance & Accounting70.763.7−7.0
Artificial Analysis Intelligence Index v5.0Indexes57.751.0−6.7
BuseyBench SVGMultimodal7.97.3−5.8
CritPt v2 (agentic, avg@5)Science82.376.9−5.4
Vibe Code Bench v1.1Code89.184.0−5.1
MMMU ProMultimodal89.284.7−4.6
LMArena Agent Arena Confirmed SuccessAgents18.49.7−4.4
VIALS-AA (pass^3)Multimodal64.060.2−3.7
AA-LCR v1.1Search79.382.7+3.3
Excel Modeling Benchmark v1.1Finance & Accounting76.573.3−3.1
FrontierCode 2.0 (Artificial Analysis)Code40.637.5−3.1
GDPxProfessional Work10.98.5−2.4
Artificial Analysis Progress Index v1Indexes351345−1.9
VIALS-AA (average of 3)Multimodal67.366.0−1.2
LMArena Agent Arena Steering BurdenAgents19.116.6−1.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work29.628.6−1.0
VIALS-AA (pass@3)Multimodal70.871.4+0.6
LMArena Agent Arena Tool HallucinationAgents0.51.0+0.2
DAXBench v2Code93.693.60.0

Gemini 4 CarbonvsGPT-6.1 Terranormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Arena and 8 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…