The Model Index
Up to four models

GPT-5.2 vs Gemini 3.1 Pro

On the index (GPT-4 = 100), GPT-5.2 scores 162 and Gemini 3.1 Pro 161. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to GPT-5.2
Presets
GPT-5.2
162146–178
WeightsClosed
ReleasedDec 2025
Price in / out$1.75 / $14
Context400K
Core tests taken6 of 9
Gemini 3.1 Pro
161146–177
WeightsClosed
ReleasedFeb 2026
Price in / out$2.0 / $12
Context1.05M
Core tests taken9 of 9

GPT-5.2 and Gemini 3.1 Pro are too close to call (0 points apart, within the uncertainty), ahead on 1 of 6 shared core tests, at a similar price.

Test by testthird-party results

GPT-5.2Gemini 3.1 Pro
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)Humanity's Last ExamHumanity's Last ExamGPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProCL-benchCL-benchEBR-benchEBR-benchSimpleBenchSimpleBenchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)Chess puzzlesChess puzzlesFrontierMath (2025 set)FrontierMath (2025 set)FrontierMath tier 4 (2025 set)FrontierMath tier 4 (2025 set)FrontierMath (tier 4)FrontierMath (tier 4)CODINGSWE-bench VerifiedSWE-bench VerifiedSWE-bench Pro (public set)SWE-bench Pro (public set)LiveCodeBenchLiveCodeBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1GSO-BenchGSO-BenchMirrorCodeMirrorCodeWeirdMLWeirdMLAGENTSDTBenchDTBenchDeepResearch BenchDeepResearch BenchFurniture assemblyFurniture assemblyLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKCorpFin v2CorpFin v2LegalBenchLegalBenchMedCodeMedCodeMedScribeMedScribeMortgageTaxMortgageTaxTaxEval v2TaxEval v2NOVEL PROBLEMSARC-AGI-2ARC-AGI-2ARC-AGI-1ARC-AGI-1
Every result as a table
TestGPT-5.2Gemini 3.1 Pro
Reasoning & knowledge
GPQA Diamond (Epoch AI)91.4%94.4%
Humanity's Last Exam27.8%46.4%
GPQA Diamond (Vals AI)91.7%95.5%
MMLU Pro86.2%91.0%
MMMU Pro86.7%88.2%
CL-bench18.2%20.8%
EBR-bench23.0%14.3%
SimpleBench45.8%79.6%
SimpleQA Verified37.1%73.5%
Maths
FrontierMath (tiers 1-3)67.4%59.7%
AIME-style maths (OTIS mock)96.1%95.6%
Chess puzzles48.0%51.0%
FrontierMath (2025 set)40.7%36.9%
FrontierMath tier 4 (2025 set)18.8%16.7%
FrontierMath (tier 4)31.7%26.8%
Coding
SWE-bench Verified73.8%75.6%
SWE-bench Pro (public set)29.9%46.1%
LiveCodeBench85.4%88.5%
SWE-bench75.8%78.8%
Vibe Code Bench v1.153.5%32.0%
GSO-Bench27.5%22.6%
MirrorCode32.2%8.9%
WeirdML72.2%72.1%
Agents
DTBench90.9%97.1%
DeepResearch Bench41.1%47.8%
Furniture assembly38.3%26.7%
LMCA43.9%53.8%
Mystery game puzzles23.0%34.0%
Professional work
CorpFin v265.9%64.5%
LegalBench82.8%87.4%
MedCode49.8%59.1%
MedScribe84.4%76.1%
MortgageTax67.1%69.4%
TaxEval v275.8%72.9%
Novel problems
ARC-AGI-272.9%77.1%
ARC-AGI-194.5%98.0%
Task length · METR45 sec46 sec

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG