Inference
EN
Anthropic

Claude Fable 5.1

Claude Fable 5.1 is a reasoning model from Anthropic in the Claude Fable family. 71 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Reasoning
  • Input: text · image · file
  • Released 2026-09
RankedEnough results for an index.
IndexAll categories, out of 100.76.2±2.4Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.69/sMedian 67/s
First tokenSeconds to the first output token.3.0 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$10.00batch $5.00Median $0.60
Output / 1MUS dollars per 1M output tokens.$50.00batch $25.00 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.1512±6 #4Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
anthropic/claude-fable-5.1
Max outputMaximum output tokens per request.
128K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.25
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
6 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Amazon BedrockCloud$10.00$50.001M
AnthropicMaker$10.00$50.001M
DigitalOceanCloud$10.00$50.001M
Google Vertex AICloud$10.00$50.001M
Microsoft AzureCloud$10.00$50.001M
OpenRouterRouter$10.00$50.001M

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.28 counted71.123 / 218
CodingCode writing and repair.18 counted77.75 / 209
ReasoningLogic problems and puzzles.10 counted75.814 / 227
MultimodalTasks with images and text.3 counted77.91 / 84
KnowledgeFacts and expert knowledge.6 counted77.26 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted74.3—
MathMath problems.5 counted77.06 / 168

Results

71 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
OTIS Mock AIME 2024-2025Math100.0%1 / 200
LiveCodeBenchCoding90.5% ±0.91 / 142
MMLU ProKnowledge92.4% ±0.31 / 135
MMMU ProMultimodal90.6% ±0.71 / 92
Artificial Analysis Omniscience AccuracyKnowledgeAbove fitFar above this model's index.68.4%2 / 164
Artificial Analysis Humanity's Last ExamKnowledge59.1%3 / 222
ProofBench v1.1Math100.0%1 / 54
LiveBench Browser UseAgentic100.0%1 / 49
Artificial Analysis Long Context ReasoningReasoning85.3%5 / 217
Artificial Analysis SciCodeCoding63.1%3 / 114
LiveBench MathematicsMath97.0%2 / 66
Artificial Analysis GPQA DiamondKnowledge93.7%7 / 199
LiveBench Coding 2.0Coding97.7%2 / 52
Humanity's Last ExamKnowledge65.0%3 / 77
LiveBench Long AgenticAgentic80.8%2 / 51
ARC-AGI-1 (semi-private)Reasoning97.5%5 / 126
GDPval-AA normalizedAgentic62.7%6 / 133
LiveBench LanguageKnowledge89.5%3 / 66
LiveBench CodingCoding86.4%3 / 66
LiveBench Agentic CodingAgentic66.1%3 / 66
Artificial Analysis Terminal-Bench v2.1Agentic91.4%1 / 21
SWE-bench ProCoding81.2%4 / 83
Vibe Code Bench v1.1Coding90.3% ±1.66 / 123
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal77.8%2 / 39
GPQA Diamond · Vals AIKnowledge93.4% ±1.97 / 135
Terminal-Bench 2.1 · Vals AIAgentic85.0% ±1.04 / 75
cursorBench32Coding73.4%1 / 18
ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobAgenticAbove fitFar above this model's index.72.0%1 / 18
Critical Physics TasksReasoningAbove fitFar above this model's index.29.7%12 / 214
Humanity's Last Exam without toolsKnowledge60.9%3 / 53
LiveBench ReasoningReasoning91.7%5 / 80
ARC-AGI-2 (semi-private)Reasoning90.0%9 / 131
MirrorCodeCoding93.3% ±3.93 / 39
LiveBench DesignAgentic83.4%4 / 52
Artificial Analysis AnalystAgentAgentic57.5%2 / 26
FrontierMath-Tiers-1-3-v2-PrivateMath90.2% ±1.87 / 90
Agent Arena command recoveryAgentic11.64 / 50
SimpleQA VerifiedKnowledge70.8% ±1.48 / 91
Agent Arena steerabilityAgentic14.85 / 54
Agent Arena task outcomeAgentic14.45 / 54
Terminal-Bench 4.0Agentic58.1% ±3.36 / 61
LiveBench Data AnalysisReasoning80.3%9 / 80
FrontierMath-Tier-4-v2-PrivateMath87.8% ±5.29 / 72
OfficeQA ProMultimodal69.0%2 / 16
Mystery Game PuzzlesReasoning58.0% ±5.010 / 80
Toolathlon-VerifiedAgentic77.8%3 / 23
IOICoding90.8% ±4.66 / 45
Code MigrationCoding54.6% ±4.811 / 77
Vibe Code Bench 1-100Coding28.0% ±4.54 / 26
ProgramBench: Can Language Models Rebuild Programs From Scratch?Coding87.6%3 / 19
SkillsBenchCoding61.6% ±4.86 / 35
OSWorld 2.1 (all 108 tasks)Agentic80.7%2 / 11
Bug Hunt BenchCoding43.0%5 / 25
MazeBench: visual spatial reasoning in a 3D open worldAgentic10.0%10 / 48
Chess PuzzlesReasoning48.0% ±5.019 / 90
Terminal-Bench 4.0.0Agentic57.9% ±1.95 / 23
Terminal-Bench 4.1.0Agentic57.3% ±2.05 / 23
EBR-benchReasoning57.1% ±6.07 / 31
Artificial Analysis Tau3-BankingAgentic47.2%5 / 22
LiveBench Instruction FollowingInstruction73.0%16 / 66
Furniture AssemblyReasoning70.0% ±5.79 / 37
Artificial Analysis ITBench-AAAgentic49.5%6 / 24
PostTrainBench v1.1Coding40.2%4 / 15
LiveBench ResearchAgentic82.0%14 / 52
FrontierSWE v2Coding56.3%6 / 22
cursorBench40Coding51.8%6 / 21
FrontierCode 2.0Coding32.4%8 / 27
Artificial Analysis GDP.pdfAgentic26.2%8 / 22
OSWorld 2.1 offline subset (82 tasks)Agentic83.5%4 / 10
Artificial Analysis Terminal-Bench v4.0Agentic52.0%10 / 22
DeepSWEAgentic67.4%29 / 59
CWE-bench v1Agentic58.0%6 / 12
Artificial Analysis DeepsecBench-AAAgentic23.5%11 / 21
OSWorld 2.0Agentic41.7%14 / 25
Artificial Analysis AutomationBenchAgentic59.4%14 / 23
Terminal-Bench-Science 0.1Agentic52.6%8 / 11
Artificial Analysis CWE-Bench-AAAgentic50.8%15 / 19
Artificial Analysis CyberGym-E2E-AAAgentic0.8%15 / 19
AutomationBenchAgentic31.4%26 / 32
Artificial Analysis Cyber IndexAgentic25.0%16 / 19
Medical Long Context Reasoning (MLCR-AA)Reasoning15.6%20 / 22
EMBFinance76.7% ±2.11 / 75
LegalBenchLegal88.5% ±0.42 / 149
ARC-AGI-2 (public eval)Reasoning98.3%2 / 122
ARC-AGI-1 (public eval)Reasoning99.0%2 / 116
MedScribeHealthcare91.3% ±2.02 / 109
MortgageTaxFinance70.8% ±0.92 / 97
Legal Research BenchLegal55.3% ±3.52 / 90
Tax Agent BenchFinance50.7% ±3.62 / 82
Public Benefits Bench v1.1Public benefits74.9% ±1.12 / 50
ProgramBenchCoding7.0% ±1.83 / 65
TaxEval v2Finance76.0% ±0.88 / 143
MedCodeHealthcare53.5% ±2.28 / 106
Finance Agent (v2)Finance58.9% ±2.18 / 79
Terminal-Bench-ScienceScience40.0% ±5.97 / 44
MysteryMechanismScience47.7% ±3.45 / 30
Time Horizon Index: KSPOther63.3%3 / 13
SRE BenchOther22.9% ±2.69 / 36
CUA-benchOther13.2%3 / 9
CyberBench v1.1Other70.4% ±5.422 / 48
FrontierMath-ErdosMath0.0%8 / 15
SpireClimb (Slay the Spire 2, Ascension 10)Other4.0% ±2.37 / 9
SpireClimbOther2.0% ±1.17 / 9
SpireClimb (Slay the Spire 1, Ascension 20 Heart)Other0.0%7 / 9
Global MMLUMultilingual94.0%—
Artificial Analysis Harvey LAB-AA v1.0Agentic93.0%—
Multi-task Indic Language Understanding BenchmarkMultilingual93.0%—
Data Research and Analysis with Complex OperationsAgentic87.7%—
Chartography with image and code toolsMultimodal86.2%—
GDP.pdf mean criteria pass rate without toolsMultimodal85.4%—
GDP.pdf mean criteria pass rate with toolsMultimodal85.1%—
BenchCAD Vision2Code voxel IoU with toolsMultimodal84.3%—
Toolathlon Verified Pass@3Agentic81.5%—
OfficeQAMultimodal80.2%—
Toolathlon Verified Pass cubedAgentic73.1%—
Artificial Analysis HealthBench Professional raw scoreKnowledge69.2%—
WANDR (Wide ANd Deep Research)Agentic68.7%—
GraphWalks BFS 256K–1MReasoning65.0%—
Artificial Analysis HealthBench raw scoreKnowledge63.9%—
AA-Briefcase v1.1, rubric pass rateAgentic61.5%—
PhysicianBenchKnowledge61.0%—
Artificial Analysis HealthBench Professional length-adjusted scoreKnowledge57.8%—
Artificial Analysis HealthBench length-adjusted scoreKnowledge57.2%—
BenchCAD 2.1 Core, with toolsMultimodal44.4%—
BenchCAD Vision2Code voxel IoU without toolsMultimodal43.7%—
Chartography without toolsMultimodal42.6%—
BenchCAD 2.1 Core, without toolsMultimodal36.3%—
Anthropic STEM Professor BenchReasoning29.9%—
RoboDojo SimAgentic27.6%—
Toolathlon Verified average assistant turnsAgentic23.7%—
Remote Labor Index (RLI)Agentic17.9%—
Harvey LAB-AA v1.1Agentic6.4%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseVals AINo licence statedEpoch AICC BYLiveBenchNo licence statedARC Prize FoundationTerms restrict reuseTerminal-BenchNo licence statedLMArenaCC BY 4.0models.devMIT

History

5 releases
Index79.0 → 76.2
23 Oct 202612 Nov 2026
Output / 1M$50.00 → $50.00
23 Oct 202612 Nov 2026

Questions

How much does Claude Fable 5.1 cost?
$10.00 per 1M input tokens, $50.00 per 1M output tokens.
How good is Claude Fable 5.1?
Its index is 76.2 out of 100. That is place 10 of 283 ranked models.
What is Claude Fable 5.1 best at?
Multimodal: 77.9 out of 100, place 1 of 84.
How fast is Claude Fable 5.1?
69 output tokens per second, 3.0 s to the first token.
What is the context window of Claude Fable 5.1?
1M tokens.
Is Claude Fable 5.1 open source?
No. Its weights are not published.
When was Claude Fable 5.1 released?
September 2026.

Badge

Claude Fable 5.1 on Inference360: place 10 of 283, index 76.2
HTML
<a href="https://inference360.ai/models/claude-fable-5-1"><img src="https://inference360.ai/badges/claude-fable-5-1.svg" alt="Claude Fable 5.1 on Inference360: place 10 of 283, index 76.2" height="20"></a>
Markdown
[![Claude Fable 5.1 on Inference360: place 10 of 283, index 76.2](https://inference360.ai/badges/claude-fable-5-1.svg)](https://inference360.ai/models/claude-fable-5-1)

Compare with

Kimi K3.1 Pro75.7GPT-6.1 Terra77.1Claude Sonnet 5.577.2Muse Wonder Full77.3GPT-5.5 Pro74.8

Lower price

GPT-6.1 Terra77.1 · $5.00/1MGemini 4 Carbon82.7 · $10.00/1MGemini 4 Argon78.5 · $10.00/1M

More from Anthropic

All models
Claude Mythos 5.1UnrankedClaude Opus 5.578.6Claude Sonnet 5.577.2Claude Haiku 5.566.3Claude Opus 573.8