The Model Index
Up to four models

GPT-5.6 Sol vs Kimi K3.1 Pro

On the index (GPT-4 = 100), GPT-5.6 Sol scores 179 and Kimi K3.1 Pro 179. 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 Sol
Presets
GPT-5.6 Sol
179164–195
WeightsClosed
ReleasedJul 2026
Price in / out$4.0 / $20
Context1.05M
Core tests taken5 of 9
Kimi K3.1 Pro
179163–195
WeightsOpen
ReleasedOct 2026
Price in / out$0.24 / $8.4
Context1.05M
Core tests taken4 of 9

GPT-5.6 Sol and Kimi K3.1 Pro are too close to call (0 points apart, within the uncertainty), ahead on 1 of 3 shared core tests, at 3.5× the price.

Test by testthird-party results

GPT-5.6 SolKimi K3.1 Pro
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)SimpleQA 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 4.0IOIIOIVibe Code Bench v1.1Vibe Code Bench v1.1AGENTSBioMysteryBenchBioMysteryBenchSRE BenchSRE BenchTerminal-Bench-ScienceTerminal-Bench-ScienceDTBenchDTBenchFurniture assemblyFurniture assemblyLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKLegal Research BenchLegal Research BenchTax Agent BenchTax Agent BenchNOVEL PROBLEMSMysteryMechanismMysteryMechanism
Every result as a table
TestGPT-5.6 SolKimi K3.1 Pro
Reasoning & knowledge
GPQA Diamond (Epoch AI)93.5%94.2%
SimpleQA Verified69.7%62.2%
Maths
FrontierMath (tiers 1-3)89.1%78.3%
AIME-style maths (OTIS mock)100.0%100.0%
ProofBench v1.183.0%91.0%
Chess puzzles59.0%45.0%
FrontierMath (tier 4)82.9%65.8%
Coding
Terminal-Bench 4.037.9%63.6%
IOI91.2%81.9%
Vibe Code Bench v1.180.5%84.9%
Agents
BioMysteryBench71.1%75.6%
SRE Bench30.5%27.1%
Terminal-Bench-Science20.0%31.4%
DTBench95.7%93.3%
Furniture assembly56.7%55.8%
LMCA58.5%57.6%
Mystery game puzzles58.0%42.0%
Professional work
Legal Research Bench48.1%37.5%
Tax Agent Bench31.1%24.9%
Novel problems
MysteryMechanism33.3%26.1%

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

This page as a chartDownload PNG