Atlas

Compare Models

Claude Sonnet 5.5 vs Kimi K3.1 Pro

Claude Sonnet 5.5 leads by 2.8 capability points (≈95% interval +0.3 to +5.3): a clear lead over Kimi K3.1 Pro · P(higher) ≈ 99%

Wins 66 of 91 shared benchmarks (23 lost, 2 tied)

Capability vs. relative cost
Claude Sonnet 5.5Kimi K3.1 ProClosed 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.

Claude Sonnet 5.5vsKimi K3.1 Procapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins66 – 23 · 91 shared, 2 tied

Comparing matching setups only (62 benchmarks excluded) · suites counted once
← Claude Sonnet 5.5Kimi K3.1 Pro →

Claude Sonnet 5.5vsKimi K3.1 Prowins by categoryData Nov 15, 2026Data: SecureBio, Epoch AI, Arena and 35 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Kimi K3.1 Pro minus Claude Sonnet 5.5.

sudo L7 (October 2026 rescoring)Code50.616.7−33.9
DAYJOB: Data ScienceProfessional Work35.34.7−30.6
HoneyBench v0.2 (hack rate)Safety41.671.8−30.2
OpenWorldEngineer (Mean@8)Code39.013.6−25.4
FrontierCode 2.0 - Pass RateCode46.422.0−24.4
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting40.517.0−23.5
Mystery Game PuzzlesGames65.042.0−23.0
MysteryMechanismScience49.126.1−23.0
FrontierCode 2.0 - ScoreCode42.219.9−22.3
DAYJOB: AccountingFinance & Accounting35.213.6−21.6
LMArena Agent Arena Net Improvement (Steering Burden)Agents16.04.9−20.9
FrontierML ($200 wallet)Agents50.429.6−20.9
DeepFinanceFinance & Accounting54.634.6−20.0
Furniture Assembly BenchmarkMultimodal75.055.8−19.2
WorkflowArenaAgents33.216.5−16.7
micro1 PharmaBenchHealth & Medicine11.427.5+16.1
Android Bench 2.0Code40.024.0−16.0
RiemannBench (November 2026 rescoring)Math76.860.8−16.0
Terminal-Bench Science 0.1 (Vals)Science47.131.4−15.7
SimpleQA Verified (Epoch anti-abstention)General QA46.562.2+15.7
WCB (SecureBio)Science66.551.5−15.0
FrontierMath Tier 4 (v2) - PrivateMath80.565.9−14.6
Bug Hunt BenchCode48.934.3−14.6
Agents' Last Exam v1.1 ScoreAgents60.474.4+14.0
Chess Puzzles (task version 2.0)Games59.045.0−14.0
Humanity's Last Exam DiamondGeneral QA53.339.4−13.9
HPCT (SecureBio)Science77.263.5−13.7
Conceptual Reasoning IndexIndexes77.167.8−12.7
Real-SWE 1.1Code51.239.2−12.0
LMArena Agent Arena Praise vs ComplaintAgents30.58.0−11.3
ALE-BenchCode24482132−11.2
LMArena Code Arena WebDevCode17731688−11.1
Visa Agentic Commerce BenchmarkAgents64.453.3−11.1
LMCAGeneral QA68.357.6−10.7
FrontierMath Tiers 1-3 (v2) - PrivateMath88.878.2−10.5
LAB-Bench ProtocolQA (SecureBio)Science88.378.1−10.2
Concept Synth INDUCTION Challenge100 Round 1General QA44.034.0−10.0
Bio Capabilities Index (SecureBio)Indexes175161−9.8
GBENCH One-shot Coding GScoreGames67.458.2−9.2
ProofBench v1.1Math100.091.0−9.0
SPU (Science Prediction Universe)Science0.30.2−8.6
LMArena Agent Arena Confirmed SuccessAgents19.52.7−8.4
ACCoRDGeneral QA85.777.5−8.2
APEX-Agents 1.1Professional Work83.775.9−7.8
AA-Briefcase v1.1 EloAgents18161678−7.6
Vibe Code Bench v1.1Code92.484.9−7.5
GameBench 2Games91.484.0−7.4
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting41.548.2+6.7
NoCoT-Bench NCKI (kspine v3)General QA107116+6.4
VCT v2 (SecureBio)Science52.246.2−6.0
DualEntry Accounting AI BenchmarkFinance & Accounting80.286.1+5.9
Excel Modeling Benchmark v1.1Finance & Accounting75.970.0−5.9
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.059.8+5.8
LiveBench 2.0Agents82.576.7−5.7
LAB-Bench CloningScenarios (SecureBio)Science97.091.5−5.5
Access to Justice BenchmarkLegal74.969.5−5.5
GPQA-Bio (SecureBio)Science87.682.2−5.4
DeepSWE v1.1Code70.575.5+5.0
Humanity's Last ExamGeneral QA46.151.1+5.0
EnterpriseBench: CoreCraftProfessional Work70.866.1−4.7
NoCoT-Bench NCRI 15.2General QA95.1103+4.3
BuseyBench SVGMultimodal6.56.9+4.1
DTBenchGeneral QA97.393.3−4.0
LMArena Agent Arena Steering BurdenAgents14.56.7−3.9
Terminal-Bench 4.0Code61.865.7+3.9
Epoch Capabilities Index (ECI)Indexes166162−3.5
Concept Synth INDUCTION Challenge64 Round 1General QA50.046.9−3.1
DAYJOB: LawLegal27.023.9−3.1
VCT (SecureBio)Science52.549.4−3.1
WMDP-Bio (SecureBio)Science92.689.7−2.9
LMArena Agent Arena Bash ReliabilityAgents12.77.2−2.8
LAB-Bench LitQA2 (SecureBio)Science55.853.1−2.7
APEX-Accounting 1.1Finance & Accounting64.562.4−2.1
MBCT (SecureBio)Science58.860.7+1.9
LMArena Text Arena EloChat & Writing14721483+1.8
MMMU ProMultimodal85.884.1−1.7
AutomationBench-AA 1.0.6Agents71.873.3+1.5
HLAB-Vals-1.1Legal10.411.7+1.3
IOI (Vals v2)Code83.181.9−1.2
MazeBench ASCIIGames1.02.0+1.0
Design Arena Overall Code CategoriesCode13491353+0.7
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting85.786.1+0.4
MineBench — Bradley-Terry RatingGames19981992−0.4
GDPval-AA v2.1Professional Work18381846+0.4
MMLU-Bio (SecureBio)Science91.091.2+0.2
Prophet Arena Arena Score (Fixed Context)Forecasting85.385.4+0.1
PubMedQA (SecureBio)Health & Medicine81.381.2−0.1
LMArena Agent Arena Tool HallucinationAgents0.91.0+0.1
DeepsecBench — Revalidated V1 (233 issues)Code37.537.50.0
DAXBench v2Code86.286.20.0
OTIS Mock AIME 2024-2025Math100.0100.00.0

Claude Sonnet 5.5vsKimi K3.1 Pronormalised score marginsData Nov 15, 2026Data: SecureBio, Epoch AI, Arena and 35 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…