Inference
EN
Anthropic

Claude Fable 5

Claude Fable 5 is a reasoning model from Anthropic in the Claude Fable family. 55 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Reasoning
  • Input: text · image · file
  • Released 2026-06
RankedEnough results for an index.
IndexAll categories, out of 100.74.0±2.7Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.56/sMedian 67/s
First tokenSeconds to the first output token.3.8 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.1492±4 #12Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
anthropic/claude-fable-5
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.
$1.00
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
7 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Amazon BedrockCloud$10.00$50.001M
Deep InfraHost$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
Snowflake CortexCloud$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.16 counted71.421 / 218
CodingCode writing and repair.14 counted74.711 / 209
ReasoningLogic problems and puzzles.9 counted73.618 / 227
MultimodalTasks with images and text.3 counted70.111 / 84
KnowledgeFacts and expert knowledge.6 counted75.88 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.2 counted64.19 / 63
MathMath problems.5 counted76.29 / 168

Results

55 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
ARC-AGI-1 (semi-private)Reasoning98.5%1 / 126
LiveCodeBenchCoding89.8% ±0.92 / 142
LiveBench LanguageKnowledge90.7%1 / 66
LiveBench Coding 2.0Coding98.3%1 / 52
τ²-Bench Tool-Agent-User EvaluationAgentic98.5%3 / 149
MMLU ProKnowledge91.5% ±0.33 / 135
Artificial Analysis Humanity's Last ExamKnowledge55.5%5 / 222
Artificial Analysis Omniscience AccuracyKnowledgeAbove fitFar above this model's index.65.4%4 / 164
Software Engineering Benchmark VerifiedCoding95.0%3 / 114
Vibe Code Bench v1.1Coding90.4% ±2.14 / 123
MMMU ProMultimodal89.3% ±0.73 / 92
Artificial Analysis SciCodeCoding61.0%6 / 114
OSWorld-VerifiedAgentic85.0%2 / 37
LiveBench ResearchAgentic86.9%3 / 52
LiveBench DesignAgentic84.7%3 / 52
LiveBench CodingCoding86.0%4 / 66
Critical Physics TasksReasoningAbove fitFar above this model's index.28.6%14 / 214
OTIS Mock AIME 2024-2025Math99.7% ±0.314 / 200
SWE-bench ProCoding80.0%6 / 83
GPQA Diamond · Vals AIKnowledge93.2% ±1.910 / 135
SWE-benchCoding95.0% ±1.07 / 87
LiveBench Browser UseAgentic92.0%4 / 49
IOI v1Coding72.3% ±9.45 / 61
ARC-AGI-2 (semi-private)Reasoning89.2%11 / 131
Artificial Analysis GPQA DiamondKnowledge92.6%17 / 199
Terminal-Bench HardAgentic62.9%2 / 23
LiveBench Instruction FollowingInstruction75.8%6 / 66
FrontierMath-Tier-4-v2-PrivateMath90.2% ±4.67 / 72
SimpleQA VerifiedKnowledge70.7% ±1.49 / 91
LiveBench Data AnalysisReasoning80.5%8 / 80
Agent Arena command recoveryAgentic9.65 / 50
FrontierCode 1.1 MainCoding53.5%2 / 17
Artificial Analysis Long Context ReasoningReasoning82.3%27 / 217
Code MigrationCoding55.1% ±4.610 / 77
FrontierMath-Tiers-1-3-v2-PrivateMath87.0% ±2.012 / 90
Terminal-Bench 2.1 · Vals AIAgentic80.5% ±1.410 / 75
GDPval-AA normalizedAgentic55.5%18 / 133
LiveBench MathematicsMath96.0%9 / 66
VulcanBench v3Coding89.5%2 / 14
LiveBench Agentic CodingAgentic62.2%10 / 66
cursorBench32Coding70.5%3 / 18
Mystery Game PuzzlesReasoning52.0% ±5.014 / 80
LiveBench Long AgenticAgentic53.2%9 / 51
MirrorCodeCoding70.0% ±10.47 / 39
ProofBench v1.1Math95.0% ±2.210 / 54
LiveBench ReasoningReasoning89.7%15 / 80
MazeBench: visual spatial reasoning in a 3D open worldAgentic11.0%9 / 48
ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobAgentic34.0%4 / 18
Terminal-Bench 4.0Agentic41.4% ±1.314 / 61
Terminal-Bench 3.0Agentic34.0%3 / 13
Terminal-Bench 2.1 · BenchLMAgentic84.3%22 / 86
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal69.8%10 / 39
Agent Arena task outcomeAgentic5.114 / 54
Chess PuzzlesReasoning47.0% ±5.024 / 90
Agent Arena steerabilityAgentic7.915 / 54
GPQA diamond · Epoch AIKnowledge85.9% ±2.569 / 217
τ²-bench BankingAgentic39.7%9 / 28
Terminal-Bench 4.1.0Agentic44.9% ±2.08 / 23
Terminal-Bench 4.0.0Agentic44.5% ±2.08 / 23
Artificial Analysis IFBenchInstruction63.5%58 / 146
FrontierSWE v2Coding47.0%9 / 22
EBR-benchReasoning39.5% ±2.013 / 31
Artificial Analysis AnalystAgentAgentic48.8%12 / 26
Artificial Analysis Terminal-Bench v2.1Agentic84.6%10 / 21
Furniture AssemblyReasoning35.8% ±6.223 / 37
OfficeQA ProMultimodal57.9%12 / 16
Artificial Analysis Tau3-BankingAgentic38.1%20 / 22
LegalBenchLegal88.6% ±0.31 / 149
CorpFin v2Finance71.8% ±0.92 / 131
FrontierMath-2025-02-28-PublicMath100.0%1 / 64
FrontierMath-Tier-4-2025-07-01-PublicMath100.0%1 / 36
TaxEval v2Finance76.9% ±0.85 / 143
MedCodeHealthcare56.1% ±2.24 / 106
EMBFinance73.7% ±2.55 / 75
ARC-AGI-2 (public eval)Reasoning96.4%9 / 122
Public Benefits Bench v1Public benefits71.7% ±1.21 / 13
MortgageTaxFinance68.9% ±0.98 / 97
Public Benefits Bench v1.1Public benefits70.4% ±1.25 / 50
Legal Research BenchLegal49.5% ±3.59 / 90
MedScribeHealthcare88.5% ±1.912 / 109
Tax Agent BenchFinance38.3% ±3.512 / 82
ARC-AGI-1 (public eval)Reasoning97.6%17 / 116
ProgramBenchCoding2.0% ±1.011 / 65
OEIS Open LiteMath44.0% ±5.01 / 5
Finance Agent (v2)Finance56.3% ±0.819 / 79
Terminal-Bench-ScienceScience15.7% ±4.413 / 44
Vals Web Search IndexOther46.9% ±2.16 / 12
FrontierMath-ErdosMath0.0%8 / 15
Artificial Analysis Harvey LAB-AA v1.0Agentic93.6%—
cursorBench31Coding70.6%—
Artificial Analysis EnterpriseOps-GymAgentic51.1%—
Blueprint-Bench 2Multimodal38.6%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYARC Prize FoundationTerms restrict reuseLiveBenchNo licence statedVals AINo licence statedTerminal-BenchNo licence statedSierra ResearchMITLMArenaCC BY 4.0models.devMIT

History

5 releases
Index76.0 → 74.0
23 Oct 202612 Nov 2026
Output / 1M$50.00 → $50.00
23 Oct 202612 Nov 2026

Questions

How much does Claude Fable 5 cost?
$10.00 per 1M input tokens, $50.00 per 1M output tokens.
How good is Claude Fable 5?
Its index is 74.0 out of 100. That is place 13 of 283 ranked models.
What is Claude Fable 5 best at?
Knowledge: 75.8 out of 100, place 8 of 268.
How fast is Claude Fable 5?
56 output tokens per second, 3.8 s to the first token.
What is the context window of Claude Fable 5?
1M tokens.
Is Claude Fable 5 open source?
No. Its weights are not published.
When was Claude Fable 5 released?
June 2026.

Badge

Claude Fable 5 on Inference360: place 13 of 283, index 74.0
HTML
<a href="https://inference360.ai/models/claude-fable"><img src="https://inference360.ai/badges/claude-fable.svg" alt="Claude Fable 5 on Inference360: place 13 of 283, index 74.0" height="20"></a>
Markdown
[![Claude Fable 5 on Inference360: place 13 of 283, index 74.0](https://inference360.ai/badges/claude-fable.svg)](https://inference360.ai/models/claude-fable)

Compare with

Claude Opus 573.8GPT-5.4 Pro73.3GPT-5.5 Pro74.8Muse Wonder73.1GPT-5.6 Sol73.0

Lower price

Hy472.0 · $1.25/1MGPT-6.1 Terra77.1 · $5.00/1MGemini 4 Carbon82.7 · $10.00/1M

More from Anthropic

All models
Claude Mythos 570.8Claude Fable LatestUnrankedClaude Opus 4.867.7Claude Sonnet 564.2Claude Haiku LatestUnranked