Atlas

Compare Models

Claude Fable 5.1 vs Grok 4.6

Claude Fable 5.1 leads by 5.7 capability points (≈95% interval +3.5 to +8.0): a clear lead over Grok 4.6 · P(higher) ≈ 100%

Wins 181 of 212 shared benchmarks (30 lost, 1 tied)

Capability vs. relative cost
Claude Fable 5.1Grok 4.6Closed 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 5.1vsGrok 4.6capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins181 – 30 · 212 shared, 1 tied

Comparing matching setups only (22 benchmarks excluded) · suites counted once
← Claude Fable 5.1Grok 4.6 →

Claude Fable 5.1vsGrok 4.6wins by categoryData Nov 15, 2026Data: Mercor, Vals AI, Artificial Analysis and 68 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.6 minus Claude Fable 5.1.

RT-SAFE Collisions per EpisodeSafety20.959.5−96.7
LLM Creative Story WritingChat & Writing3.8-2.9−75.0
Document ArenaProfessional Work15151438−68.6
AA-PersonalityBench Imagination indexChat & Writing86.720.0−67.3
KernelBench Mega Normalized (RTX PRO 6000)Code64.70.0−64.7
AA-PersonalityBench Warmth indexChat & Writing94.331.5−63.5
Medical Long Context Reasoning (MLCR-AA)Health & Medicine71.112.2−58.9
Time Horizon Index: KSPAgents63.35.8−57.5
FrontierMath Tier 4 (v2) - PrivateMath87.831.7−56.1
AA-PersonalityBench Informativeness indexChat & Writing96.244.8−51.9
Gray Swan IPI Attack Success (k=15)Safety1.051.8−50.8
ProofBench v1.1Math100.051.0−49.0
AA-PersonalityBench Organization indexChat & Writing60.014.3−46.2
LMArena Agent Arena Net ImprovementAgents14.11.1−42.7
BioSecBench-Refusal V1 Clean RerunsSafety5.646.9+41.3
InferenceBenchCode9.86.7−37.9
ALE-BenchCode28491822−36.4
Antidote: Everyday EditionSafety1065966−35.4
AA-PersonalityBench Expressiveness indexChat & Writing98.163.8−34.6
CheatBenchSafety45.879.1−33.3
Terminal-Bench 4.0Code57.925.3−32.6
MirrorCode (ML, +Private, 2L)Code73.341.1−32.2
Humanity's Last Exam DiamondGeneral QA51.319.6−31.7
CAD ArenaAgents66.234.8−31.4
Hemingway-benchChat & Writing1095983−31.0
Furniture Assembly BenchmarkMultimodal70.040.0−30.0
Terminal-Bench Science 0.1 (Vals)Science34.34.3−30.0
HardGeoBench Avg DistanceMultimodal8451807−29.7
LMArena Agent Arena Net Improvement (Steering Burden)Agents14.8-0.2−28.3
AA-PersonalityBench Transparency indexChat & Writing98.171.4−26.9
Terminal-Bench 2.1 Extended (Mercor held-out)Code19.545.5+26.0
ChartographyMultimodal46.220.5−25.7
WeirdML (v2)Science92.967.3−25.6
Arena Image-to-WebDevCode17241607−25.4
GameBench 2Games94.068.7−25.3
HardGeoBench <=10miMultimodal45.020.0−25.0
FrontierMath Tiers 1-3 (v2) - PrivateMath90.266.0−24.2
BuseyBench SVGMultimodal7.04.6−24.2
Mystery Game PuzzlesGames58.034.0−24.0
ARC-AGI-2General QA90.067.1−22.9
NoCoT-Bench NCRI 15.2General QA12886.8−22.7
MMMU-Pro Extended (Mercor held-out)Multimodal47.626.0−21.6
SimpleQA Verified (Epoch anti-abstention)General QA70.849.3−21.5
HardGeoBench <=100miMultimodal53.832.5−21.2
Humanity's Last ExamGeneral QA60.939.7−21.2
CloudOpsBench v1.0Agents60.639.7−20.9
HealthBench ProfessionalHealth & Medicine69.248.5−20.7
HackerRank-BenchCode47.527.7−19.8
SAGEMath48.528.9−19.6
SciUniverse Level 1Science45.326.2−19.1
LLM Debate BenchmarkChat & Writing17381604−19.0
SWE Atlas Codebase QnA Extended (Mercor held-out)Code38.019.3−18.7
MedXpertQA MM Extended (Mercor held-out)Health & Medicine66.048.4−17.6
MysteryMechanismScience47.730.6−17.1
Conceptual Reasoning IndexIndexes74.962.5−17.1
LMArena Agent Arena Praise vs ComplaintAgents31.6-2.5−17.1
LMCAGeneral QA65.548.5−17.0
SWE-bench Verified Extended (Mercor held-out)Code61.378.0+16.7
Design Creativity BenchCode48.643.4−16.7
AA-AnalystAgent (pass^5)Professional Work57.541.3−16.2
LMArena Code Arena WebDevCode17421618−16.2
Roboflow Vision Evals (October 2026, five tasks)Multimodal74.759.1−15.6
ReactBench (39 tasks)Code49.233.8−15.4
Artificial Analysis Coding Agent Index v1.5Indexes62.247.0−15.3
Chess Puzzles (task version 2.0)Games48.033.0−15.0
HardGeoBench Country AccuracyMultimodal58.843.8−15.0
Toloka Arena (v3)Agents71.056.0−15.0
CritPt v2 (agentic, avg@5)Science76.662.3−14.3
GBENCH One-shot Coding GScoreGames69.355.4−14.0
Excel Modeling BenchmarkFinance & Accounting76.762.7−13.9
KernelBench CUDA Normalized (RTX PRO 6000)Code51.637.7−13.9
SWE-Marathon v1.1Code45.631.9−13.8
Bug Hunt BenchCode41.027.3−13.7
Vals RSI Index v1.1 (historical five tasks)Agents35.021.5−13.5
LiveBench 2.0Agents88.074.9−13.0
Roboflow Vision Evals (September 2026 protocol)Multimodal81.368.7−12.6
Real-SWECode45.032.5−12.5
ChessBench EloGames1095891−12.1
ACCoRDGeneral QA83.771.6−12.1
SWE-bench VerifiedCode96.684.6−12.0
EnterpriseBench: CoreCraftProfessional Work77.465.6−11.8
AA-PersonalityBench Slop indexChat & Writing23.835.3+11.5
NoCoT-Bench NCKI (kspine v3)General QA128112−11.5
HANDBOOK.mdProfessional Work38.827.3−11.5
CritPt (avg@5)Science31.119.7−11.4
RSI-Exam 0.1 (Full 88)Agents0.50.4−11.4
GeneBench-Pro Public Release SubsetScience26.015.0−11.0
Toloka Arena (v1)Agents60.749.7−11.0
AutoResearchExam (hidden-test AUARC)Agents0.60.5−10.9
SimpleBenchGeneral QA86.675.9−10.7
Vals RSI Index v1.1 (four tasks)Agents36.125.5−10.6
FrontierCode 2.0 (Artificial Analysis)Code30.319.8−10.5
CursorBench 4.0Code51.841.4−10.4
Finance Agent v2 - PrecedentsFinance & Accounting32.221.9−10.3
ARC-AGI-1 Semi-PrivateGeneral QA97.587.5−10.0
Extended NYT Connections 940 (Quadratic v1)Games90.080.0−10.0
ProgramBench (Mercor 100-task subset)Code11.31.7−9.6
SciCode (main problems)Science24.214.6−9.6
Immigration BenchLegal75.466.0−9.5
BioSecBench-Structure V1 (58 evaluations)Science6.315.5+9.2
Eval Connections One-Shot with TrapsGames82.091.0+9.0
SUPERChemScience74.765.7−9.0
Artificial Analysis Intelligence Index v4.3Indexes53.444.4−9.0
Snorkel SWE-bench CLICode14.55.7−8.8
MedCodeHealth & Medicine53.544.7−8.8
Blueprint-Bench 2Multimodal0.40.3−8.7
SWE-TogetherCode69.360.6−8.7
ComplexConstraintsChat & Writing45.136.5−8.6
LMArena Agent Arena Confirmed SuccessAgents14.4-2.6−8.5
MedXpertQA MM (200-question set)Health & Medicine83.775.2−8.5
Senior SWE-BenchCode34.726.3−8.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.045.7−8.3
LMArena Agent Arena Steering BurdenAgents14.8-1.8−8.3
LMArena Text Arena EloChat & Writing15031456−8.1
Concept Synth INDUCTION Challenge100 Round 1General QA33.025.0−8.0
WellnessBenchSafety51.043.0−8.0
MBABench 2.0Finance & Accounting7.90.0−7.9
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science3.611.5+7.9
AA-LCR Extended (Mercor held-out)Search73.065.2−7.8
LibraryDesignBenchCode47.539.7−7.8
AA-PersonalityBench Conviction indexChat & Writing10092.4−7.7
Epoch Capabilities Index (ECI)Indexes167158−7.4
ChessBench CoherenceGames61.969.2+7.3
Finance Agent v2 - Market AnalysisFinance & Accounting69.562.2−7.3
Design Arena Overall Code CategoriesCode13311288−7.1
MineBench — Bradley-Terry RatingGames19281811−7.1
DAYJOB: FinanceFinance & Accounting19.812.8−7.0
Shroombench (mushroom identification, top-1)Multimodal51.044.0−7.0
SWE Atlas - Codebase QnACode43.350.3+7.0
Biopharma Bench V0.1Health & Medicine69.963.0−6.9
GDPval Extended (Mercor held-out rubric score)Professional Work68.761.8−6.9
DAYJOB: HealthcareHealth & Medicine9.62.8−6.8
Terminal-Bench 2.1Code81.688.4+6.8
Vision Arena OverallMultimodal12891263−6.7
SciCode 1.0.1 (Artificial Analysis)Code63.156.5−6.6
MortgageTax (Vals all 1,258 documents)Finance & Accounting70.864.2−6.6
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting79.773.2−6.5
AA-LCR v1.1Search87.581.0−6.5
GlobeBenchGeneral QA21.628.1+6.5
Tax Agent BenchFinance & Accounting77.971.4−6.5
Finance Agent v2 - Financial ModelingFinance & Accounting25.318.9−6.4
APEX-Accounting 1.1Finance & Accounting66.760.3−6.4
Epoch Automation ReportsAgents65.359.1−6.2
SnorkelUnderwrite 2.0Finance & Accounting24.017.8−6.2
APEX-1 (Mercor, mean score)Professional Work73.867.7−6.1
AA-Briefcase EloAgents16501545−6.1
Snorkel Agentic Coding 2.0Code39.633.6−6.0
Finance Agent v2 - Disclosure AnalysisFinance & Accounting70.364.4−5.9
APEX-SWECode63.657.7−5.9
BrowseComp (Mercor 130-question subset)Search78.384.0+5.7
ATLAS-Finance v1.1Finance & Accounting10.34.7−5.7
MMMU ProMultimodal86.180.5−5.6
SnorkelFinance 2.0Finance & Accounting19.625.1+5.5
Olam Social Poker EloGames14941499+5.4
DAXBench v2Code91.586.2−5.3
SnorkelManufacturingProfessional Work11.316.4+5.1
CharXiv Extended (Mercor held-out)Multimodal82.277.3−4.9
Finance Agent v2 - All PassFinance & Accounting47.842.9−4.9
TaxEval v2Finance & Accounting76.071.1−4.9
SnorkelLegalLegal35.940.7+4.8
Finance Agent v2 - AdjustmentsFinance & Accounting54.449.6−4.8
MedScribeHealth & Medicine91.386.5−4.8
Finance Agent v2 - ComparablesFinance & Accounting46.041.2−4.8
RuneBenchGames6.15.8−4.8
CVDP (517-problem set)Code62.658.0−4.6
GDPval-AA v2Professional Work17451663−4.0
BabaIsBench Stage 0: The IntroGames100.096.0−4.0
CadQueryEval (Corrected Geometry)Code76.072.0−4.0
ChessBench AccuracyGames71.367.8−3.8
GDPval-AA v2.1Professional Work17411658−3.6
SpeciEval (Pooled English Runs)Safety87.290.8+3.6
SpeciEval (Refusal-Aware Retries, Historical)Safety87.290.8+3.6
MMLU ProGeneral QA93.089.4−3.6
AdvancedIFChat & Writing84.280.9−3.3
Realm Financial ReasoningFinance & Accounting47.844.6−3.2
Snorkel Digital Work IndexIndexes23.620.5−3.1
LMArena Agent Arena Bash RecoveryAgents11.65.7−3.0
Clipping Bench v1.0Games1068805−2.8
LMArena Agent Arena SteerabilityAgents8.43.3−2.6
SWEeper-BenchCode56.559.0+2.5
LiveCodeBench (Vals AI public dataset)Code90.588.2−2.3
FrontierCode 1.1 Extended - ScoreCode63.661.3−2.3
SAB ResearchFinance & Accounting96.894.6−2.2
LegalBenchLegal88.586.3−2.2
CharXiv (descriptive and reasoning combined)Multimodal95.192.9−2.2
ChessBenchGames45.343.3−2.0
SnorkelRevOpsProfessional Work14.015.8+1.8
FrontierCode 1.1 Extended - Pass RateCode68.867.0−1.7
Finance Agent v2 - Earnings AnalysisFinance & Accounting75.573.9−1.6
SWE-bench MultilingualCode91.192.7+1.6
Concept Synth INDUCTION Challenge64 Round 1General QA34.435.9+1.6
GPQA DiamondGeneral QA93.995.5+1.5
TxBench-Antibody-Development V1 (60 evaluations)Science26.928.3+1.4
Snorkel WorkplaceAgentsProfessional Work15.817.1+1.3
CWE-bench v1 Programmatic Pass@1Cybersecurity58.057.0−1.0
OTIS Mock AIME 2024-2025Math100.099.2−0.8
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting77.077.8+0.8
ObviousBenchGeneral QA99.3100.0+0.7
Arena Alignment IndexSafety80.180.8+0.6
DeepSWE v1.1Code67.366.7−0.6
GDPval (Mercor 213-task rubric score)Professional Work86.686.0−0.6
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.085.6−0.4
Hedge-BenchFinance & Accounting26.826.4−0.4
micro1 PharmaBenchHealth & Medicine12.812.5−0.3
DTBenchGeneral QA97.697.3−0.3
APEX-Agents 1.1Professional Work77.177.3+0.2
LMArena Agent Arena Tool HallucinationAgents0.81.0+0.1
HieroglyphBenchMultimodal20.420.3−0.1
Prophet Arena Arena Score (Fixed Context)Forecasting85.085.0−0.1
SM BenchSafety89.089.00.0
LMArena Agent Arena Bash ReliabilityAgents10.110.00.0
AIME 2025Math100.0100.00.0

Claude Fable 5.1vsGrok 4.6normalised score marginsData Nov 15, 2026Data: Mercor, Vals AI, Artificial Analysis and 68 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…