Benchmark Almanac

Claude Fable 5

Claude Fable 5 ranks first on 39 of the 245 benchmarks with five or more models scored.

Anthropic Released June 9th, 2026 Family Claude 5 The lab's model card Compare with others
253
Benchmarks with a score
39
Ranked first, out of five or more models
82
In the top three
Anthropic Jun 2026

Claude Fable 5

Closed weights Flagship

164.1 #11 of 256
Epoch Capabilities Index
Self-Reported Index 162.1 #20 of 404
  • DeepSWE 69.9% #5 /28
  • AutomationBench 17.1% #19 /42
  • Agents' Last Exam 25.7% #15 /31

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
EQ-Bench 3 2,096 #1 of 78 Official leaderboard
LMArena (Elo) 1,504 #5 of 335 Official leaderboard
Hemingway-bench 1,110 #1 of 42 Official leaderboard
Creative Writing v3 81.9 #3 of 120 Official leaderboard
Document Arena 1,500 #1 of 39 Official leaderboard
Antidote 1,099 #1 of 36 Official leaderboard
Longform Creative Writing (EQ-Bench) 83.0 #4 of 119 Official leaderboard
GraphWalks 99.96% #1 of 27 Lab model card
Humanity's Last Exam 64.5% #7 of 154 Lab model card
BullshitBench V2 83.6% #7 of 129 Official leaderboard
DesignArena ASCII Art 1,321 #5 of 77 Official leaderboard
HUMAINE 34 #4 of 50 Official leaderboard
SimpleQA Verified 70.7% #9 of 94 Epoch AI run
Lech Mazur Creative Writing 2.81 #6 of 50 Official leaderboard
HLE Extended 58.7% #4 of 31 Official leaderboard
StudyArena Text 1,280 #11 of 85 Official leaderboard
EQ-Bench 4 1,313 #5 of 32 Official leaderboard
Extended NYT Connections 92.7% #17 of 103 Official leaderboard
AA-LCR Extended 71.1% #12 of 49 Official leaderboard

Math and science

BenchmarkScoreRankSource
ARC-AGI 98.5% #1 of 95 Official leaderboard
OTIS Mock AIME 2024-2025 100% #3 of 212 Epoch AI run
DTBench 98.4% #3 of 180 Official leaderboard
ProofBench 77% #2 of 59 Official leaderboard
Surface Evolver Bench 95% #1 of 25 Official leaderboard
ForecastBench 0.12 #5 of 82 Official leaderboard
CritPt 28.6% #1 of 13 Lab model card
SimpleBench 81.9% #10 of 106 Official leaderboard
FrontierMath Tier 4 v2 90.2% #7 of 68 Official leaderboard
ARC-AGI-2 89.2% #11 of 98 Official leaderboard
xbench-ScienceQA 73.4% #4 of 29 Official leaderboard
FrontierMath Tiers 1-3 v2 87% #12 of 86 Official leaderboard
ProofBench v1.1 95% #9 of 52 Official leaderboard
Into the Unknown 0% #2 of 9 Official leaderboard
Verifying the Verifiers (Project APE CRED benchmark) 96.5% #6 of 23 Official leaderboard
PaperBenchX 11.8% #3 of 10 Official leaderboard
GPQA Diamond 85.9% #78 of 229 Epoch AI run
Terminal-Bench-Science 21.4% #6 of 17 Official leaderboard
BioMysteryBench 83.9% #5 of 10 Lab model card
ArXivMath 78.5% #5 of 9 Lab model card
FrontierMath Erdős 0% #9 of 15 Official leaderboard
Prophet Arena 84.9 #21 of 24 Official leaderboard
Riemann-bench 55% #4 of 4 Lab model card

Coding

BenchmarkScoreRankSource
SWE-bench Verified 95.5% #2 of 155 Lab model card
OpenHands Index 81% #1 of 33 Official leaderboard
WeirdML 92% #4 of 112 Official leaderboard
Vibe Code Bench 90.3% #5 of 120 Official leaderboard
FrogsGame post-training task 34% #1 of 17 Official leaderboard
FrontierSWE v1 88.2% #1 of 17 Official leaderboard
SWE-Bench Pro 80.3% #5 of 81 Lab model card
AndroidBench 90.9% #2 of 32 Official leaderboard
DesignArena Agentic Game Dev 1,233 #2 of 27 Official leaderboard
APEX-SWE 58.9% #5 of 65 Official leaderboard
SWE-rebench 64.5% #1 of 13 Official leaderboard
DesignArena Android (Kotlin) 1,276 #4 of 46 Official leaderboard
ALE-Bench 2,463 #11 of 126 Official leaderboard
IOI (Vals) 72.3% #5 of 55 Official leaderboard
SWE-bench Verified Extended 78.7% #4 of 44 Official leaderboard
SWE-Together 68.8% #2 of 22 Official leaderboard
GBAEval 74.5% #2 of 21 Official leaderboard
DesignArena Image to Frontend 1,264 #3 of 31 Official leaderboard
GSO-Bench 78.4% #3 of 30 Official leaderboard
Vision2Web 82.3% #1 of 10 Official leaderboard
DesignArena SVG 1,294 #11 of 103 Official leaderboard
DesignArena React Native 1,257 #6 of 56 Official leaderboard
DesignArena Data Visualization 1,304 #14 of 118 Official leaderboard
DesignArena Game Dev 1,326 #17 of 143 Official leaderboard
WebDev Arena 1,628 #13 of 105 arena.ai, via Epoch AI
DesignArena Godot Game Dev 1,274 #3 of 24 Official leaderboard
FrontierCode 1.1 53.5% #1 of 8 Official leaderboard
Code Migration 55.1% #10 of 78 Official leaderboard
PostTrainBench 30.7% #3 of 23 Official leaderboard
Senior SWE-Bench 34.7% #3 of 23 Official leaderboard
SWE Atlas Test Writing 55.6% #3 of 22 Official leaderboard
ReactBench 44.6% #4 of 28 Official leaderboard
DesignArena UI Component 1,311 #20 of 139 Official leaderboard
DesignArena Code 1,302 #22 of 143 Official leaderboard
DesignArena Agentic Overall 1,266 #12 of 74 Official leaderboard
DesignArena Website 1,289 #26 of 151 Official leaderboard
SWE Atlas Refactoring 54.8% #3 of 17 Official leaderboard
DeepSWE 69.9% #5 of 28 Official leaderboard
MirrorCode 70% #7 of 39 Epoch AI run
Image-to-WebDev Arena 1,621 #10 of 55 Official leaderboard
PostTrainBench 1.1 41.8% #2 of 11 Official leaderboard
DesignArena Full Stack 1,259 #9 of 49 Official leaderboard
DesignArena 3D Design 1,299 #25 of 136 Official leaderboard
Code Arena 1,625 #24 of 128 Official leaderboard
MLS-Bench 49.9% #3 of 15 Official leaderboard
CursorBench 3.1 72.9% #1 of 5 Lab model card
QwenQoderBench 63.1% #1 of 5 Official leaderboard
QwenReactBench 1,770 #1 of 5 Official leaderboard
QwenSWEBench 86.3% #1 of 5 Official leaderboard
Z.ai Code Bench 39.5% #1 of 5 Official leaderboard
GameDevBench 67.3% #4 of 17 Official leaderboard
SWE-Marathon 24% #3 of 12 Official leaderboard
GameCraft-Bench 65.7% #4 of 15 Official leaderboard
DesignArena Frontend Web App 1,243 #15 of 53 Official leaderboard
QwenBlenderBench 69.5% #2 of 6 Official leaderboard
DesignArena Image to HTML 1,217 #20 of 51 Official leaderboard
FrontierSWE 47% #10 of 25 Official leaderboard
WeirdML v3 15.6% #11 of 27 Official leaderboard
PostTrainBench 1.2 37.3% #9 of 21 Official leaderboard
Vals RSI Index 0.24 #12 of 27 Official leaderboard
OpenRSI Index Learnability-Aware Long/Short CoT Adaptation 47.5% #1 of 2 Official leaderboard
OpenRSI Index Molmo2 Video-Pointing Inference Strategy 0.46 #1 of 2 Official leaderboard
SWE Atlas 39% #12 of 22 Official leaderboard
QwenSVGBench 1,690 #3 of 5 Official leaderboard
SWE Atlas Codebase QnA Extended 0% #38 of 38 Official leaderboard
MiMo Visual Coding 69.1% #9 of 9 Official leaderboard
FrontierCode 29.3% #4 of 4 Lab model card

Agents and tools

BenchmarkScoreRankSource
BrowseComp 93.3% #1 of 81 Lab model card
Pencil Puzzle Bench 97.6% #1 of 53 Official leaderboard
OSWorld-Verified 86% #1 of 33 Official leaderboard
Toolathlon 61.7% #1 of 29 Lab model card
BabaIsBench 100% #2 of 32 Official leaderboard
LisanBench 13680.3 #10 of 133 Official leaderboard
PACT 1,603 #2 of 26 Official leaderboard
CEO-Bench $12,630,078 #2 of 19 Official leaderboard
WildClawBench 62% #4 of 31 Official leaderboard
LMArena Search Arena 1,229 #5 of 37 Official leaderboard
RuneBench 6.01 #11 of 76 Official leaderboard
Terminal-Bench 2.1 83.8% #2 of 13 Official leaderboard
Mystery Game Puzzles 52% #14 of 84 Epoch AI run
ApprenticeBench 43% #3 of 18 Official leaderboard
Agent Arena 7.30% #11 of 56 Official leaderboard
BrowseComp Extended 44.5% #10 of 42 Official leaderboard
DuelLab GameBench 2 75.2 #14 of 57 Official leaderboard
Chess Puzzles 47% #24 of 93 Epoch AI run
MCP Atlas 83.3% #10 of 37 Official leaderboard
Terminal-Bench 3.0 34% #3 of 11 Official leaderboard
Terminal-Bench 4.0 44.5% #11 of 40 Official leaderboard
Terminal-Bench 4.1 44.9% #12 of 43 Official leaderboard
Long-Horizon-Terminal-Bench (LHTB) 48.7% #7 of 25 Official leaderboard
Bench on the Clocktower 52.7% #3 of 10 Official leaderboard
GameHorizon 71.2 #13 of 37 Official leaderboard
τ³-bench Banking 39.7% #9 of 25 Official leaderboard
EBR-bench 39.5% #13 of 31 Epoch AI run
AutomationBench 17.1% #19 of 42 Official leaderboard
MCPMark Verified 86.6% #4 of 8 Official leaderboard
Clocktower Bench 1,544 #6 of 11 Official leaderboard
Vals Web Search Index 48.5% #3 of 5 Official leaderboard
PinchBench 2.0 54.8% #44 of 57 Official leaderboard
Terminal-Bench 2.1 Extended 16.8% #39 of 44 Official leaderboard
TextQuests 23.5 #25 of 28 Official leaderboard

Professional work

BenchmarkScoreRankSource
LegalBench 88.6% #1 of 138 Official leaderboard
CorpFin v2 71.8% #2 of 119 Official leaderboard
TaxEval v2 76.9% #5 of 129 Official leaderboard
MedCode 56.1% #4 of 96 Official leaderboard
GDPval-AA 1,932 #1 of 16 Lab model card
Excel Modeling Benchmark 73.7% #5 of 76 Official leaderboard
3DCodeArena 1,230 #1 of 14 Official leaderboard
OfficeQA Pro 60.9% #1 of 12 Official leaderboard
RadLE 2.0 758.0 #1 of 12 Official leaderboard
TRLawBench 9.64 #3 of 34 Official leaderboard
CADGenBench 45.1% #1 of 11 Official leaderboard
MortgageTax 68.9% #8 of 87 Official leaderboard
Vals Index v2 66% #6 of 65 Official leaderboard
EnterpriseBench: CoreCraft 70.3% #5 of 51 Official leaderboard
Public Benefits Bench 70.4% #5 of 50 Official leaderboard
E-Commerce Bench 8.05 #2 of 19 Official leaderboard
Legal Research Bench 49.5% #10 of 93 Official leaderboard
MedScribe 88.5% #12 of 97 Official leaderboard
PRBench 52.6% #6 of 48 Official leaderboard
JobBench 57.4% #2 of 16 Official leaderboard
HealthBench Professional 66% #4 of 26 Lab model card
Vals Index v2.1 61.4% #8 of 49 Official leaderboard
Tax Agent Bench 38.3% #14 of 84 Official leaderboard
DesignArena Python-PPTX Slides 1,252 #6 of 36 Official leaderboard
Harvey LAB (Legal Agent Benchmark) 11.3% #14 of 80 Official leaderboard
YC-Bench $1,977,573 #11 of 59 Official leaderboard
HealthBench 62.7% #6 of 31 Lab model card
CoWorkBench 75.9% #1 of 5 Official leaderboard
DesignArena HTML Slides 1,173 #6 of 28 Official leaderboard
prinzbench 56.0 #6 of 26 Official leaderboard
GDPval Extended 2.90% #10 of 43 Official leaderboard
Vals Finance Agent v2 56.3% #19 of 80 Official leaderboard
Tuesday Work Index 61.5 #11 of 44 Official leaderboard
APEX-Agents 1.1 74.5% #18 of 65 Official leaderboard
APEX-Agents 63.6% #17 of 61 Collected by Epoch AI, source not named
OfficeQA Pro V2 45.6% #2 of 7 Official leaderboard
DAYJOB: Healthcare 11.2% #11 of 37 Official leaderboard
Genspark Slides Benchmark 79.9% #3 of 10 Official leaderboard
DAYJOB: Finance 18.3% #12 of 39 Official leaderboard
Remote Labor Index 15.8% #6 of 19 Official leaderboard
APEX-1 66% #23 of 71 Official leaderboard
Physical-Engineering's Last Exam 0.34 #6 of 18 Official leaderboard
Real-World Finance (Anthropic) 70% #1 of 3 Lab model card
CHI-Bench 24% #8 of 23 Official leaderboard
APEX-Accounting 61% #18 of 49 Official leaderboard
FrontierFinance 49.2% #7 of 17 Official leaderboard
Vending-Bench 2 $5,680 #28 of 67 Official leaderboard
Management Consulting Tasks (OpenAI) 35.5% #3 of 7 Official leaderboard
LifeSci Diamond 3.60% #8 of 18 Official leaderboard
MedXpertQA MM Extended 56.9% #19 of 41 Official leaderboard
PolicyBench 83.3% #24 of 50 Official leaderboard
Agents' Last Exam 25.7% #15 of 31 Official leaderboard
RedlineBench 47.3% #2 of 4 Official leaderboard
TaxCalcBench v2 46% #17 of 32 Official leaderboard
EEBench 54.3% #16 of 30 Official leaderboard
Vending-Bench Arena $4,244 #18 of 32 Official leaderboard
Real-World Finance v2 (Anthropic) 1,374 #2 of 3 Lab model card
OfficeQA 57.9% #13 of 14 Lab model card

Vision, audio and robotics

BenchmarkScoreRankSource
Vision Arena 1,309 #1 of 136 Official leaderboard
CharXiv 93.5% #1 of 75 Lab model card
GeoBench 89% #1 of 48 Official leaderboard
BabyVision 90.5% #2 of 27 Official leaderboard
ZeroBench 46% #7 of 76 Official leaderboard
MineBench 1,851 #13 of 75 Official leaderboard
VoxelBench 2,112 #10 of 56 Official leaderboard
PerceptionBench (Moonshot) 57.2% #3 of 16 Official leaderboard
ArtifactArena 1,189 #4 of 21 Official leaderboard
EnigmaEval 39.3% #1 of 5 Official leaderboard
AgenticVBench 32.4% #2 of 9 Official leaderboard
Chartography 34.8% #10 of 43 Official leaderboard
ParseBench 69.8% #11 of 44 Official leaderboard
Blueprint-Bench 2 38.6% #10 of 38 Official leaderboard
Drone-Bench 84.3% #5 of 19 Official leaderboard
MMMU-Pro Extended 37.6% #13 of 38 Official leaderboard
Humanity's Sixth Sense 34.5% #12 of 34 Official leaderboard
ClockBench 37.2% #14 of 37 Official leaderboard
CharXiv Extended 80.7% #18 of 38 Official leaderboard
GDP.pdf 29.8% #10 of 17 Lab model card
Furniture Assembly Benchmark 35.8% #22 of 35 Official leaderboard
QwenVisualOffice 32.4% #5 of 6 Official leaderboard
GlobeBench 7.29% #18 of 19 Official leaderboard

Safety and security

BenchmarkScoreRankSource
nocot-bench (NCRI) 119.0 #8 of 185 Official leaderboard
Dictatorship Eval 91.3% #1 of 20 Official leaderboard
Single-turn benign request evaluations (Anthropic) 0.01% #3 of 19 Lab model card
Anthropic automated behavioral audit 2.06 #1 of 6 Lab model card
Last Secure Code Benchmark 25.1% #1 of 6 Official leaderboard
Refusal to assist with AI safety R&D (Anthropic) 3.10% #1 of 5 Lab model card
Gray Swan Agent Red Teaming (ART) benchmark 0.10% #4 of 14 Lab model card
PCT (Polarized Contrastive Pairs) 39.8% #14 of 46 Official leaderboard
Malicious use of Claude Code (Anthropic) 90.3% #6 of 17 Lab model card
KORA 53.8% #16 of 42 Official leaderboard
Prompt injection robustness against adaptive attackers, coding (Anthropic) 0.45% #7 of 14 Lab model card
Anthropic undermining liberal democracy (behavioral audit metric) 1.20 #3 of 6 Lab model card
Lying about identity (Anthropic) 99% #1 of 2 Lab model card
BBQ (Bias Benchmark for QA) 84.5% #14 of 25 Lab model card
Anthropic political even-handedness eval 97.6% #8 of 14 Lab model card
Malicious computer use (Anthropic) 85.7% #8 of 14 Lab model card
DistressBench 69.4% #26 of 42 Official leaderboard
Prompt injection robustness against adaptive attackers, computer use (Anthropic) 0.82% #9 of 14 Lab model card
Rewind-attacker cyber adversarial robustness (Anthropic) 5.00% #4 of 6 Lab model card
Single-turn harmful request evaluations (Anthropic) 96.9% #13 of 19 Lab model card
Prompt injection robustness against adaptive attackers, browser use (Anthropic) 29.7% #12 of 13 Lab model card

Other

BenchmarkScoreRankSource
ACCoRD 85.0 #4 of 152 Official leaderboard
Vals Multimodal Index 74.2% #1 of 33 Official leaderboard
Conceptual Reasoning Index 72.7 #8 of 151 Official leaderboard
LMCA 61.1% #9 of 147 Official leaderboard
CAIS Risk Index 30.8 #5 of 67 Official leaderboard
SAGE (Vals) 51.9% #9 of 81 Official leaderboard
CAIS Text Capabilities Index 54.4 #8 of 68 Official leaderboard
CAIS Vision Capabilities Index 66.6 #10 of 53 Official leaderboard

Its lab's numbers put it at 162.1, 2.0 points below its ECI

The Self-Reported Index fits one score to the 21 capability results Anthropic reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 159.2 to 165.0. How we compute it.

Its lab's numbers next to 9 leaderboards

Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.

Benchmark Lab says Board says Gap
Harvey LAB (Legal Agent Benchmark) 16.9% Best of Legal Legal Agent Benchmark (Harvey's Held-Out Set) / Legal Agent Full Public Benchmark Set 11.3% Leaderboard +5.6 pts
Terminal-Bench 2.1 88% Agentic coding Terminal-Bench 2.1 83.8% Leaderboard, xhigh, Claude Code +4.2 pts
OfficeQA Pro 57.9% OfficeQA Pro, Databricks' own evaluation with documents read as images (Table 8.1.A, 8.17.1) 60.9% Leaderboard, Claude Agent SDK −3.0 pts
OSWorld-Verified 85% Computer use OSWorld-Verified 86% Leaderboard, 1M context, 100 steps −1.0 pts
AutomationBench 17.4% Tool use AutomationBench 17.1% Leaderboard, max +0.3 pts
Blueprint-Bench 2 38.6% Spatial reasoning Blueprint-Bench 2 38.6% Leaderboard Same
MCP Atlas 83.3% MCP-Atlas pass rate (8.17.5) 83.3% Leaderboard Same
Vals Finance Agent v2 56.3% Finance Agent Benchmark v2, run by Vals AI, adaptive thinking at max effort (8.17.2) 56.3% Leaderboard Same
Vending-Bench 2 $5,680 Vending-Bench 2 final balance at high effort, Fable 5's best effort level (corrected from max effort in the June 11 changelog) (8.17.6) $5,680 Leaderboard, high Same

Anthropic's models over time

Anthropic's models in color, against every other model on the Epoch Capabilities Index, by release date.

50 100 150 2023 2024 2025 2026 2027 Claude Fable 5.2 Claude Haiku 5.5 Claude Sonnet 4.6 Claude Haiku 4.5 Claude 3.7 Sonnet Claude 3.5 Haiku Claude 3.5 Sonnet (October 2024) 50 100 150 2023 2024 2025 2026 2027 Claude Fable 5.2 Claude Haiku 5.5

Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

Claude Fable 5 and Claude Mythos 5 (launch post); System Card: Claude Fable 5 & Claude Mythos 5, Table 8.1.A Jun 9, 2026

Its system card

Every result in the longer documents the lab published with it. 31 of them aren't in the launch post.

System Card: Claude Fable 5 & Claude Mythos 5 Jun 9, 2026

Capabilities

Safety

Robustness

Bias

Alignment

Persuasion