Atlas

Compare Models

Gemini 4 Carbon vs GPT-5.5

Gemini 4 Carbon leads by 7.9 capability points (≈95% interval +5.5 to +10.3): a clear lead over GPT-5.5 · P(higher) ≈ 100%

Wins 43 of 46 shared benchmarks (2 lost, 1 tied)

Capability vs. relative cost
Gemini 4 CarbonGPT-5.5Closed weightsOpen weights

Benchmark wins43 – 2 · 46 shared, 1 tied

Comparing matching setups only (14 benchmarks excluded) · suites counted once
← Gemini 4 CarbonGPT-5.5 →
Data: Epoch AI, Artificial Analysis, ARC Prize Foundation and 16 othersatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-5.5 minus Gemini 4 Carbon.

ARC-AGI-3 (Semi-Private)Games99.10.4−98.7
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering60.65.1−55.6
HieroglyphBenchPerception & media › Documents & charts76.022.8−53.2
ChartographyPerception & media › Documents & charts72.531.0−41.5
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge69.127.9−41.2
AutomationBench 1.0.6 (Zapier)Tool & computer use › Customer service & operations53.016.6−36.4
Furniture Assembly BenchmarkPerception & media › Spatial & visual reasoning77.544.2−33.3
TxBench Antibody Discovery V1 (100 evaluations)Science › Life sciences65.033.0−32.0
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations77.847.3−30.5
EnterpriseBench: CoreCraftTool & computer use › Customer service & operations80.551.3−29.2
AA-Briefcase v1.1 EloProfessional › General knowledge work16111087−28.7
RiemannBenchMath80.055.2−24.8
LLM Debate BenchmarkLanguage & writing › Language & wordplay16761507−24.2
SimpleBenchKnowledge & reasoning › General knowledge92.669.0−23.6
FrontierMath Tier 4 (v2) - PrivateMath95.172.5−22.6
Blueprint-Bench 2Perception & media › Spatial & visual reasoning0.60.4−22.4
TxBench Oligo Discovery V1 (113 evaluations)Science › Life sciences55.833.9−21.8
APEX-Agents 1.1Professional › General knowledge work90.470.3−20.1
DrugDiscoveryBenchScience › Life sciences71.551.6−19.9
Vibe Code Bench v1.1Software › Frontend & design89.169.8−19.2
BuseyBench SVGPerception & media › Generation7.96.0−18.6
GDPval-AA v2.1Professional › General knowledge work17331335−17.5
Mystery Game PuzzlesGames73.056.0−17.0
Tax Agent BenchProfessional › Finance, accounting & tax76.460.5−16.0
Legal Research BenchProfessional › Legal55.840.4−15.4
Chess Puzzles (task version 2.0)Games78.063.0−15.0
SimpleQA Verified (Epoch anti-abstention)Knowledge & reasoning › General knowledge77.763.0−14.7
APEX-Accounting 1.1Professional › Finance, accounting & tax70.756.3−14.4
DeepsecBench — Revalidated V1 (233 issues)Software › Cybersecurity35.221.0−14.2
Artificial Analysis Progress Index v1Composite indexes351320−10.8
TxBench Preclinical Pharmacology V1 Clean RerunsScience › Life sciences66.055.3−10.7
HiL-Bench (Human-in-Loop Benchmark)Tool & computer use › General agents50.339.7−10.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine19.428.9+9.4
ARC-AGI-2Knowledge & reasoning › Abstract reasoning94.285.0−9.2
MMMU ProKnowledge & reasoning › General knowledge89.281.2−8.1
LMArena Text Arena EloLanguage & writing › General assistance15261481−7.8
Epoch Capabilities Index (ECI)Composite indexes170160−7.6
GBENCH One-shot Coding GScoreGames69.061.6−7.4
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work29.622.6−7.0
GPQA DiamondScience › Multidisciplinary96.590.7−5.9
DAXBench v2Software › Data & ML93.688.3−5.3
FrontierMath Tiers 1-3 (v2) - PrivateMath90.585.3−5.3
AA-LCR v1.1Knowledge & reasoning › Reading & long context79.384.3+5.0
ARC-AGI-1 Semi-PrivateKnowledge & reasoning › Abstract reasoning98.595.0−3.5
FrontierMath ErdősMath2.90.0−2.9
OTIS Mock AIME 2024-2025Math100.0100.00.0
Data: Epoch AI, Artificial Analysis, ARC Prize Foundation and 16 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…