Atlas

Compare Models

Claude Fable 5.1 vs Kimi K3.1 Pro

Claude Fable 5.1 leads by 1.9 capability points (≈95% interval −0.5 to +4.3): statistically tied · P(higher) ≈ 94%

Wins 58 of 83 shared benchmarks (22 lost, 3 tied)

Capability vs. relative cost
Claude Fable 5.1Kimi 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 Fable 5.1vsKimi K3.1 Procapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins58 – 22 · 83 shared, 3 tied

Comparing matching setups only (32 benchmarks excluded) · suites counted once
← Claude Fable 5.1Kimi K3.1 Pro →

Claude Fable 5.1vsKimi K3.1 Prowins by categoryData Nov 15, 2026Data: Epoch AI, Arena, Surge AI and 38 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 Fable 5.1.

SPU (Science Prediction Universe)Science-0.50.2+35.6
ALE-BenchCode28492132−25.4
CloudOpsBench v1.0Agents43.819.1−24.7
ArXivMath 08/2026 (Native Harnesses)Math87.764.9−22.8
Visa Agentic Commerce BenchmarkAgents75.653.3−22.2
FrontierMath Tier 4 (v2) - PrivateMath87.865.9−22.0
MysteryMechanismScience47.726.1−21.6
DAYJOB: Data ScienceProfessional Work26.04.7−21.3
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting36.517.0−19.5
LMArena Agent Arena Net Improvement (Steering Burden)Agents14.84.9−18.7
Mystery Game PuzzlesGames58.042.0−16.0
FrontierCode 2.0 - Pass RateCode37.622.0−15.6
OpenWorldEngineer (Mean@8)Code29.113.6−15.5
sudo L7 (October 2026 rescoring)Code32.216.7−15.5
BrokenArXiv 08/2026 (Native Harnesses)Math79.864.3−15.5
micro1 PharmaBenchHealth & Medicine12.827.5+14.7
Terminal-Bench-Science 0.1Science52.638.1−14.5
Furniture Assembly BenchmarkMultimodal70.055.8−14.2
FrontierCode 2.0 - ScoreCode34.019.9−14.1
Tax Agent BenchFinance & Accounting77.963.9−14.0
Humanity's Last Exam DiamondGeneral QA51.337.3−14.0
NoCoT-Bench NCRI 15.2General QA128103−13.8
DAYJOB: AccountingFinance & Accounting27.213.6−13.6
Concept Synth INDUCTION Challenge64 Round 1General QA34.446.9+12.5
FrontierML ($200 wallet)Agents41.929.6−12.3
FrontierMath Tiers 1-3 (v2) - PrivateMath90.278.2−11.9
LMArena Agent Arena Praise vs ComplaintAgents31.68.0−11.8
EnterpriseBench: CoreCraftProfessional Work77.466.1−11.3
LiveBench 2.0Agents88.076.7−11.2
GBENCH One-shot Coding GScoreGames69.358.2−11.1
DeepFinanceFinance & Accounting45.634.6−11.0
AutoRLBench v2 (solution level)Agents42.152.6+10.5
RiemannBench (November 2026 rescoring)Math71.260.8−10.4
GameBench 2Games94.084.0−10.0
HoneyBench v0.2 (hack rate)Safety61.871.8−9.9
Conceptual Reasoning IndexIndexes74.967.8−9.8
ProofBench v1.1Math100.091.0−9.0
SimpleQA Verified (Epoch anti-abstention)General QA70.862.2−8.6
NoCoT-Bench NCKI (kspine v3)General QA128116−8.3
DeepSWE v1.1Code67.375.5+8.2
LMCAGeneral QA65.557.6−7.9
Terminal-Bench 4.0Code57.965.7+7.8
LMArena Code Arena WebDevCode17421688−7.1
Excel Modeling Benchmark v1.1Finance & Accounting76.770.0−6.7
Bug Hunt BenchCode41.034.3−6.7
Real-SWE 1.1Code45.639.2−6.4
ACCoRDGeneral QA83.777.5−6.2
SWE-Marathon - release-reported version unspecifiedCode45.651.7+6.1
RWE-benchAgents7.01.0−6.0
LMArena Agent Arena Confirmed SuccessAgents14.42.7−5.8
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.059.8+5.8
Longform Creative WritingChat & Writing85.379.9−5.4
DAXBench v2Code91.586.2−5.3
Humanity's Last ExamGeneral QA46.551.1+4.6
APEX-Accounting 1.1Finance & Accounting66.762.4−4.3
DTBenchGeneral QA97.693.3−4.3
Epoch Capabilities Index (ECI)Indexes167162−4.1
LMArena Agent Arena Steering BurdenAgents14.86.7−4.0
MineBench — Bradley-Terry RatingGames19281992+3.9
Design Arena Overall Code CategoriesCode13311353+3.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal17.413.8−3.6
GDPval-AA v2.1Professional Work17651846+3.6
LMArena Text Arena EloChat & Writing15031483−3.4
Chess Puzzles (task version 2.0)Games48.045.0−3.0
Artificial Analysis Intelligence Index v5.0Indexes49.046.0−3.0
DAYJOB: LawLegal26.523.9−2.6
MMMU ProMultimodal86.184.1−2.0
Creative Writing v3 Rubric ScoreChat & Writing16.316.6+1.8
LMArena Agent Arena Bash ReliabilityAgents10.17.2−1.4
Android Bench 2.0Code22.724.0+1.3
APEX-Agents 1.1Professional Work77.175.9−1.2
Concept Synth INDUCTION Challenge100 Round 1General QA33.034.0+1.0
DualEntry Accounting AI BenchmarkFinance & Accounting85.186.1+1.0
BuseyBench SVGMultimodal7.06.9−0.9
Creative Writing v3 EloChat & Writing21552173+0.9
AA-Briefcase v1.1 EloAgents16711678+0.4
Prophet Arena Arena Score (Fixed Context)Forecasting85.085.4+0.4
LMArena Agent Arena Tool HallucinationAgents0.81.0+0.1
BenchCAD - geometric overlapMultimodal92.692.7+0.1
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.086.10.0
AutoRLBench v2 (issue level)Agents25.025.00.0
MazeBench ASCIIGames2.02.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0

Claude Fable 5.1vsKimi K3.1 Pronormalised score marginsData Nov 15, 2026Data: Epoch AI, Arena, Surge AI and 38 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…