The Model Index
Up to four models

GPT-5.6 Luna vs Gemini 3.1 Pro

On the index (GPT-4 = 100), GPT-5.6 Luna 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-5.6 Luna
Presets
GPT-5.6 Luna
162146–178
WeightsClosed
ReleasedJul 2026
Price in / out$0.20 / $1.2
Context1.05M
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-5.6 Luna and Gemini 3.1 Pro are too close to call (1 point apart, within the uncertainty), ahead on 4 of 6 shared core tests, and costs 10× less.

Test by testthird-party results

GPT-5.6 LunaGemini 3.1 Pro
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)GPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProSimpleBenchSimpleBenchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)ProofBench v1.1ProofBench v1.1Chess puzzlesChess puzzlesFrontierMath (tier 4)FrontierMath (tier 4)CODINGTerminal-Bench 4.0Terminal-Bench 4.0Code MigrationCode MigrationIOIIOIProgramBenchProgramBenchSWE-benchSWE-benchVibe Code Bench 1-100Vibe Code Bench 1-100Vibe Code Bench v1.1Vibe Code Bench v1.1DeepSWEDeepSWEMirrorCodeMirrorCodeWeirdMLWeirdMLAGENTSAPEX-AgentsAPEX-AgentsBioMysteryBenchBioMysteryBenchSRE BenchSRE BenchTerminal-Bench-ScienceTerminal-Bench-ScienceBALROGBALROGDTBenchDTBenchFurniture assemblyFurniture assemblyLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)CorpFin v2CorpFin v2EMBEMBLegalBenchLegalBenchLegal Research BenchLegal Research BenchMedCodeMedCodeMedScribeMedScribeMortgageTaxMortgageTaxPublic Benefits Bench v1.1Public Benefits Bench v1.1Tax Agent BenchTax Agent BenchTaxEval v2TaxEval v2NOVEL PROBLEMSARC-AGI-2ARC-AGI-2ARC-AGI-1ARC-AGI-1ARC-AGI-3ARC-AGI-3MysteryMechanismMysteryMechanism
Every result as a table
TestGPT-5.6 LunaGemini 3.1 Pro
Reasoning & knowledge
GPQA Diamond (Epoch AI)91.6%94.4%
GPQA Diamond (Vals AI)91.7%95.5%
MMLU Pro86.0%91.0%
MMMU Pro85.0%88.2%
SimpleBench46.8%79.6%
SimpleQA Verified41.0%73.5%
Maths
FrontierMath (tiers 1-3)82.1%59.7%
AIME-style maths (OTIS mock)98.3%95.6%
ProofBench v1.160.0%26.0%
Chess puzzles43.0%51.0%
FrontierMath (tier 4)61.0%26.8%
Coding
Terminal-Bench 4.011.6%2.5%
Code Migration44.5%17.3%
IOI61.8%51.8%
ProgramBench0.0%0.0%
SWE-bench93.0%78.8%
Vibe Code Bench 1-10022.6%6.7%
Vibe Code Bench v1.177.1%32.0%
DeepSWE67.2%11.7%
MirrorCode21.1%8.9%
WeirdML60.9%72.1%
Agents
APEX-Agents43.0%35.1%
BioMysteryBench61.5%60.0%
SRE Bench1.5%0.0%
Terminal-Bench-Science0.0%1.4%
BALROG45.6%57.0%
DTBench89.1%97.1%
Furniture assembly42.5%26.7%
LMCA48.5%53.8%
Mystery game puzzles21.0%34.0%
Professional work
Finance Agent (v2)55.0%43.0%
CorpFin v264.2%64.5%
EMB67.1%52.6%
LegalBench84.0%87.4%
Legal Research Bench36.5%20.7%
MedCode42.4%59.1%
MedScribe84.4%76.1%
MortgageTax67.3%69.4%
Public Benefits Bench v1.161.2%53.8%
Tax Agent Bench24.8%9.3%
TaxEval v276.2%72.9%
Novel problems
ARC-AGI-259.6%77.1%
ARC-AGI-190.7%98.0%
ARC-AGI-30.2%0.4%
MysteryMechanism14.4%20.7%
Task length · METR46 sec

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

This page as a chartDownload PNG