Atlas

Compare Models

Claude Fable 5.1 vs DeepSeek V4.1 Flash 1017

Claude Fable 5.1 leads by 4.8 capability points (≈95% interval +2.3 to +7.3): a clear lead over DeepSeek V4.1 Flash 1017 · P(higher) ≈ 100%

Wins 88 of 97 shared benchmarks (9 lost)

Capability vs. relative cost
Claude Fable 5.1DeepSeek V4.1 Flash 1017Closed 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 5.1vsDeepSeek V4.1 Flash 1017capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins88 – 9 · 97 shared

Comparing matching setups only (13 benchmarks excluded) · suites counted once
← Claude Fable 5.1DeepSeek V4.1 Flash 1017 →

Claude Fable 5.1vsDeepSeek V4.1 Flash 1017wins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Surge AI and 34 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is DeepSeek V4.1 Flash 1017 minus Claude Fable 5.1.

HardGeoBench Avg DistanceMultimodal8453201−72.7
AA-PersonalityBench Warmth indexChat & Writing94.327.7−67.3
AA-PersonalityBench Imagination indexChat & Writing86.737.2−50.0
HardGeoBench <=100miMultimodal53.87.5−46.2
AA-PersonalityBench Expressiveness indexChat & Writing98.152.4−46.2
ArXivMath 08/2026 (Native Harnesses)Math87.743.9−43.9
HardGeoBench Country AccuracyMultimodal58.815.0−43.7
FrontierMath Tier 4 (v2) - PrivateMath87.846.3−41.5
HardGeoBench <=10miMultimodal45.03.8−41.2
LisanBench Difficulty-WeightedGames45737733+33.8
BrokenArXiv 08/2026 (Native Harnesses)Math79.846.7−33.0
RiemannBench (November 2026 rescoring)Math71.238.4−32.8
ALE-BenchCode28492012−29.7
AA-PersonalityBench Organization indexChat & Writing60.088.6+28.8
SimpleQA Verified (Epoch anti-abstention)General QA70.842.5−28.3
Terminal-Bench 4.0Code57.931.0−26.9
BuseyBench SVGMultimodal7.04.3−26.8
NoCoT-Bench NCRI 15.2General QA12881.6−25.5
CritPt v2 (agentic, avg@5)Science77.752.3−25.4
DAYJOB: Healthcare (October 2026 rescoring)Health & Medicine31.26.0−25.2
AA-PersonalityBench Informativeness indexChat & Writing96.271.4−25.0
Arena Image-to-WebDevCode17241609−24.9
LMArena Agent Arena Net Improvement (Steering Burden)Agents14.82.0−24.2
GameBench 2Games94.069.9−24.1
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting36.512.5−24.0
HANDBOOK.md Agents v1.1Professional Work47.124.0−23.1
Build-A-Chip-Bench (20-task index run)Agents23.50.5−23.0
DAYJOB: Data ScienceProfessional Work26.03.3−22.7
DAYJOB: LawLegal26.56.2−20.3
DAYJOB: AccountingFinance & Accounting27.27.6−19.6
FrontierMath Tiers 1-3 (v2) - PrivateMath90.271.2−18.9
EnterpriseBench: CoreCraftProfessional Work77.458.5−18.9
LisanBench Path LengthGames1510021107+18.7
NoCoT-Bench NCKI (kspine v3)General QA128102−18.5
FrontierML ($200 wallet)Agents41.923.5−18.4
Humanity's Last ExamGeneral QA60.942.6−18.3
Visa Agentic Commerce BenchmarkAgents75.657.8−17.8
GBENCH One-shot Coding GScoreGames69.353.8−15.5
AA-PersonalityBench Slop indexChat & Writing23.839.1+15.4
LMArena Agent Arena Praise vs ComplaintAgents31.61.2−15.2
ReactBench (39 tasks)Code49.234.4−14.9
Excel Modeling BenchmarkFinance & Accounting76.761.8−14.9
FrontierCode 2.0 - Pass RateCode37.623.0−14.6
HoneyBench v0.2 (hack rate)Safety61.876.4−14.6
APEX-SWECode63.649.1−14.5
FrontierCode 2.0 - ScoreCode34.020.2−13.8
DAXBench v2Code91.577.7−13.8
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.041.6−12.4
Concept Synth INDUCTION Challenge100 Round 1General QA33.021.0−12.0
ARC-AGI-2General QA90.078.1−11.9
AA-PersonalityBench Transparency indexChat & Writing98.186.7−11.5
LMArena Code Arena WebDevCode17421664−10.3
Mystery Game PuzzlesGames58.049.0−9.0
AA-Briefcase v1.1 EloAgents15771416−8.8
LiveBench 2.0Agents88.079.2−8.8
DAYJOB: TaxFinance & Accounting9.61.2−8.4
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine18.326.7+8.3
FrontierPhysics v0.1Science11.33.6−7.7
Excel Modeling Benchmark v1.1Finance & Accounting76.769.2−7.6
APEX-Accounting 1.1Finance & Accounting66.759.3−7.4
SciCode 1.0.1 (Artificial Analysis)Code63.155.9−7.2
CritPt (avg@5)Science31.124.0−7.1
Design Arena Overall Code CategoriesCode13311289−7.0
Finance Agent v2 - Market AnalysisFinance & Accounting69.562.7−6.9
MMMU ProMultimodal86.179.7−6.4
HealthBench Professional (Length-Adjusted)Health & Medicine58.151.7−6.4
Epoch Capabilities Index (ECI)Indexes167159−6.2
LMArena Agent Arena Confirmed SuccessAgents14.42.3−6.0
AA-PersonalityBench Conviction indexChat & Writing10094.3−5.8
Finance Agent v2 - PrecedentsFinance & Accounting32.226.5−5.7
APEX-1 (Mercor, mean score)Professional Work73.868.3−5.5
Android Bench 2.0Code22.717.3−5.4
LMArena Agent Arena Steering BurdenAgents14.84.0−5.4
FrontierCode 1.1 Main (Artificial Analysis)Code49.043.7−5.3
AA-LCR v1.1Search87.582.7−4.8
CursorBench 4.0Code51.847.7−4.1
Finance Agent v2 - Disclosure AnalysisFinance & Accounting70.366.3−4.0
Ruby LLM Benchmarks — Program FixerCode76.072.0−4.0
Finance Agent v2 - AdjustmentsFinance & Accounting54.458.4+4.0
LMArena Agent Arena Bash ReliabilityAgents10.12.4−3.9
Bug Hunt BenchCode41.037.3−3.7
GDPxProfessional Work8.55.4−3.1
ARC-AGI-1 Semi-PrivateGeneral QA97.594.5−3.0
Tax Agent BenchFinance & Accounting77.974.9−3.0
DualEntry Accounting AI BenchmarkFinance & Accounting85.182.2−2.9
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting77.074.1−2.9
Finance Agent v2 - Financial ModelingFinance & Accounting25.322.8−2.6
Finance Agent v2 - All PassFinance & Accounting47.845.2−2.5
APEX-Agents 1.1Professional Work77.179.4+2.3
Chess Puzzles (task version 2.0)Games48.046.0−2.0
Finance Agent v2 - ComparablesFinance & Accounting46.047.5+1.6
Concept Synth INDUCTION Challenge64 Round 1General QA34.432.8−1.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal17.418.8+1.4
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting79.779.2−0.6
OTIS Mock AIME 2024-2025Math100.099.4−0.6
LMArena Agent Arena Tool HallucinationAgents0.80.3−0.3
Finance Agent v2 - Earnings AnalysisFinance & Accounting75.575.3−0.2

Claude Fable 5.1vsDeepSeek V4.1 Flash 1017normalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Surge AI and 34 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…