Atlas

Compare Models

Claude Opus 5.5 vs Kimi K3.1 Pro

Claude Opus 5.5 leads by 3.9 capability points (≈95% interval +1.4 to +6.4): a clear lead over Kimi K3.1 Pro · P(higher) ≈ 100%

Wins 86 of 109 shared benchmarks (19 lost, 4 tied)

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

Benchmark wins86 – 19 · 109 shared, 4 tied

Comparing matching setups only (40 benchmarks excluded) · suites counted once
← Claude Opus 5.5Kimi K3.1 Pro →

Claude Opus 5.5vsKimi K3.1 Prowins by categoryData Nov 15, 2026Data: SecureBio, Epoch AI, Arena and 43 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 Opus 5.5.

HoneyBench v0.2 (hack rate)Safety31.671.8−40.1
ArXivMath 10/2026 (Native Harnesses)Math79.845.7−34.1
DAYJOB: Data ScienceProfessional Work37.34.7−32.6
Concept Synth INDUCTION Challenge100 Round 1General QA65.034.0−31.0
ArXivLean 09/2026Math47.217.0−30.2
FrontierCode 2.0 - Pass RateCode51.822.0−29.8
FrontierML ($200 wallet)Agents58.829.6−29.3
FrontierMath Tier 4 (v2) - PrivateMath95.065.9−29.1
Mystery Game PuzzlesGames71.042.0−29.0
sudo L7 (October 2026 rescoring)Code45.616.7−28.9
CloudOpsBench v1.0Agents47.819.1−28.7
OpenWorldEngineer (Mean@8)Code41.613.6−28.0
RWE-benchAgents29.01.0−28.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents19.54.9−27.6
Furniture Assembly BenchmarkMultimodal83.355.8−27.5
FrontierCode 2.0 - ScoreCode47.319.9−27.4
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting42.517.0−25.5
SPU (Science Prediction Universe)Science-0.30.2+25.1
Concept Synth INDUCTION Challenge64 Round 1General QA71.946.9−25.0
Chess Puzzles (task version 2.0)Games67.045.0−22.0
ALE-BenchCode27482132−21.8
CWE-bench v2 Pass@1Cybersecurity67.145.3−21.8
ArXivMath 08/2026 (Native Harnesses)Math86.064.9−21.0
DAYJOB: AccountingFinance & Accounting34.413.6−20.8
Terminal-Bench-Science 0.1Science58.738.1−20.6
WorkflowArenaAgents37.016.5−20.5
DeepFinanceFinance & Accounting54.834.6−20.2
BrokenArXiv 10/2026 (Native Harnesses)Math77.858.0−19.8
RiemannBench (November 2026 rescoring)Math80.060.8−19.2
TCSExam LiteMath72.353.5−18.8
LAB-Bench LitQA2 (SecureBio)Science71.153.1−18.0
ArXivMath 09/2026 (Native Harnesses)Math82.164.2−17.9
FrontierMemory v1.1Agents73.355.6−17.8
Real-SWE 1.1Code56.839.2−17.6
micro1 PharmaBenchHealth & Medicine10.827.5+16.7
LMArena Code Arena WebDevCode18141688−16.6
LMArena Agent Arena Praise vs ComplaintAgents40.38.0−16.2
LAB-Bench ProtocolQA (SecureBio)Science94.078.1−15.9
GBENCH One-shot Coding GScoreGames73.958.2−15.7
Humanity's Last Exam DiamondGeneral QA54.639.4−15.2
VCT v2 (SecureBio)Science60.546.2−14.3
Conceptual Reasoning IndexIndexes77.967.8−13.8
Android Bench 2.0Code37.324.0−13.3
BrokenArXiv 09/2026 (Native Harnesses)Math76.763.5−13.2
BrokenArXiv 08/2026 (Native Harnesses)Math77.464.3−13.1
FrontierMath Tiers 1-3 (v2) - PrivateMath91.278.2−13.0
VCT (SecureBio)Science61.449.4−12.0
Artificial Analysis Intelligence Index v5.0Indexes58.046.0−12.0
ACCoRDGeneral QA88.477.5−10.9
DeepsecBench — Revalidated V1 (233 issues)Code26.637.5+10.9
GameBench 2Games94.884.0−10.7
LMCAGeneral QA68.257.6−10.6
SimpleQA Verified (Epoch anti-abstention)General QA72.262.2−10.0
LiveBench 2.0Agents86.576.7−9.8
DAXBench v2Code95.786.2−9.6
ProofBench v1.1Math100.091.0−9.0
LMArena Agent Arena Confirmed SuccessAgents20.72.7−9.0
Visa Agentic Commerce BenchmarkAgents62.253.3−8.9
GPQA-Bio (SecureBio)Science90.982.2−8.7
MineBench — Bradley-Terry RatingGames21341992−8.6
EnterpriseBench: CoreCraftProfessional Work74.466.1−8.3
MBCT (SecureBio)Science68.760.7−8.0
τ³-Banking (Fireworks, banking_knowledge)Agents56.063.3+7.3
Bio Capabilities Index (SecureBio)Indexes171161−6.8
AA-Briefcase v1.1 EloAgents18011678−6.7
PubMedQA (SecureBio)Health & Medicine87.981.2−6.7
WMDP-Bio (SecureBio)Science96.189.7−6.4
Frontier3DBenchCode62.356.0−6.3
Creative Writing v3 EloChat & Writing20512173+6.2
LMArena Agent Arena Steering BurdenAgents18.66.7−5.9
Excel Modeling Benchmark v1.1Finance & Accounting75.870.0−5.8
DTBenchGeneral QA98.993.3−5.6
SWE-Marathon - release-reported version unspecifiedCode46.351.7+5.4
APEX-Agents 1.1Professional Work81.375.9−5.4
Bug Hunt BenchCode39.734.3−5.4
Epoch Capabilities Index (ECI)Indexes169162−5.4
ZeroBench Main Mean pass@1 (five samples)Multimodal35.029.8−5.2
Design Arena Overall Code CategoriesCode13841353−5.1
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.059.8+4.8
Longform Creative WritingChat & Writing84.679.9−4.7
Access to Justice BenchmarkLegal73.869.5−4.3
LMArena Text Arena EloChat & Writing15061483−4.0
ZeroBench (Main questions, pass@5)Multimodal43.039.0−4.0
LMArena Agent Arena Bash ReliabilityAgents15.17.2−4.0
DualEntry Accounting AI BenchmarkFinance & Accounting82.286.1+3.9
BenchCAD - geometric overlapMultimodal96.292.7−3.5
Creative Writing v3 Rubric ScoreChat & Writing16.016.6+3.2
DeepSWE v1.1Code72.375.5+3.2
APEX-Accounting 1.1Finance & Accounting65.562.4−3.1
Epiq LAER CorporateBenchProfessional Work75.478.4+3.1
WCB (SecureBio)Science54.551.5−3.0
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.813.8+3.0
HLAB-Vals-1.1Legal9.211.7+2.5
MMMU ProMultimodal86.684.1−2.5
Humanity's Last Exam (with tools)General QA67.770.0+2.3
MMLU-Bio (SecureBio)Science92.791.2−1.5
Chartography (with tools)Multimodal89.090.2+1.2
DAYJOB: LawLegal25.123.9−1.2
Legacy-Bench (Fireworks, avg@3)Code62.063.0+1.0
BuseyBench SVGMultimodal6.86.9+0.9
GDPval-AA v2.1Professional Work18661846−0.9
Terminal-Bench 4.0Code64.865.7+0.9
LMArena Agent Arena Tool HallucinationAgents0.71.0+0.2
Prophet Arena Arena Score (Fixed Context)Forecasting85.585.4−0.1
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.286.1−0.1
Agents' Last Exam v1.1 (ALE v1.1)Agents43.443.40.0
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting48.248.20.0
Harvey LAB: In-House Contracting (held-out)Legal3.03.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0

Claude Opus 5.5vsKimi K3.1 Pronormalised score marginsData Nov 15, 2026Data: SecureBio, Epoch AI, Arena and 43 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…