Inference360
Anthropic

Claude Fable 5

Claude Fable 5 is a reasoning model from Anthropic in the Claude Fable family. 53 benchmarks count toward its score, in 7 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.77.5 ±2.7
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.34/s
Input / 1MUS dollars per 1M input tokens.$10 batch $5
Output / 1MUS dollars per 1M output tokens.$50 batch $25 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1493 (#7)

50 is the middle of the board. The range shows the doubt in the index. Batch work costs less.

30,057 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
74.4
CodingCode writing and repair.
77.7
ReasoningLogic problems and puzzles.
79.4
MultimodalTasks with images and text.
71.4
KnowledgeFacts and expert knowledge.
77.5
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
66.5
MathMath problems.
80.6

Results

53 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
OTIS Mock AIME 2024-2025Math99.7%67.3max effort—Epoch AI
τ²-Bench Tool-Agent-User EvaluationAgentic98.5%69.5——Victor Barres et al.
ARC-AGI-1 (semi-private)Reasoning98.5%73.7max effort—ARC Prize Foundation
LiveBench MathematicsMath96.0%76.2max effort25 Jun 2026LiveBench
Software Engineering Benchmark VerifiedCoding95.0%73.7——Carlos E. Jimenez et al.
ProofBench v1.1Math95.0%87.9—21 Sept 2026Vals AI
SWE-benchCoding95.0%73.7—1 Sept 2026Vals AI
Artificial Analysis Harvey LAB-AAAgentic93.6%77.6——Artificial Analysis
GPQA DiamondKnowledge93.2%64.3—1 Sept 2026Vals AI
Artificial Analysis GPQA DiamondKnowledge92.6%63.5——Artificial Analysis
MMLU ProKnowledge91.5%64.7—1 Sept 2026Vals AI
LiveBench LanguageKnowledge90.7%79.9max effort25 Jun 2026LiveBench
Vibe Code Bench v1.1Coding90.4%79.8OpenHands21 Sept 2026Vals AI
FrontierMath-Tier-4-v2-PrivateMath90.2%91.2max effort—Epoch AI
LiveCodeBenchCoding89.8%65.9—1 Sept 2026Vals AI
LiveBench ReasoningReasoning89.7%80.5max effort25 Jun 2026LiveBench
VulcanBench v3Coding89.5%76.7——VulcanBench contributors
MMMU ProMultimodal89.3%72.7—1 Sept 2026Vals AI
ARC-AGI-2 (semi-private)Reasoning89.2%84.7max effort—ARC Prize Foundation
FrontierMath-Tiers-1-3-v2-PrivateMath87.0%80.4max effort—Epoch AI
LiveBench CodingCoding86.0%80.4max effort25 Jun 2026LiveBench
GPQA diamondKnowledge85.9%57.5max effort—Epoch AI
OSWorld-VerifiedAgentic85.0%72.9——Tianbao Xie et al.
Artificial Analysis Long Context ReasoningReasoning82.3%65.2——Artificial Analysis
LiveBench Data AnalysisReasoning80.5%67.8max effort25 Jun 2026LiveBench
Terminal-Bench 2.1Agentic80.5%71.6—21 Sept 2026Vals AI
SWE-bench ProCoding80.0%81.4——Xiang Deng et al.
Artificial Analysis Coding IndexCoding76.5%72.9——Artificial Analysis
LiveBench Instruction FollowingInstruction75.8%79.5max effort25 Jun 2026LiveBench
IOI v1Coding72.3%81.0—9 Aug 2026Vals AI
SimpleQA VerifiedKnowledge70.7%87.0xhigh effort—Epoch AI
cursorBench31Coding70.6%———Benchmark authors
cursorBench32Coding70.5%78.4——Benchmark authors
Artificial Analysis Omniscience AccuracyKnowledge65.4%94.8——Artificial Analysis
Medical Long Context Reasoning (MLCR-AA)Reasoning64.4%94.2——Wisedocs and Artificial Analysis
MirrorCodeCoding63.9%—high effort—Epoch AI
Artificial Analysis IFBenchInstruction63.5%53.5——Artificial Analysis
Terminal-Bench HardAgentic62.9%———Artificial Analysis
LiveBench Agentic CodingAgentic62.2%75.2max effort25 Jun 2026LiveBench
Artificial Analysis SciCodeCoding61.0%77.5——Artificial Analysis
OfficeQA ProMultimodal57.9%70.2——OfficeQA Pro authors
Artificial Analysis Humanity's Last ExamKnowledge55.5%84.7——Artificial Analysis
Code MigrationCoding55.1%80.2—21 Sept 2026Vals AI
GDPval-AA normalizedAgentic54.8%77.7——Artificial Analysis
FrontierCode 1.1 MainCoding53.5%81.2——Cognition
Mystery Game PuzzlesReasoning52.0%89.2max effort—Epoch AI
Artificial Analysis EnterpriseOps-GymAgentic51.1%77.3——Artificial Analysis
Artificial Analysis Agentic IndexAgentic51.0%78.7——Artificial Analysis
Artificial Analysis Intelligence IndexKnowledge49.6%84.5——Artificial Analysis
Artificial Analysis AnalystAgentAgentic48.8%75.8——Artificial Analysis
FrontierSWE v2Coding47.0%80.4——Proximal
OEIS Open LiteMath44.0%—high effort—Epoch AI
Chess PuzzlesReasoning41.0%76.5high effort—Epoch AI
τ²-bench BankingAgentic39.7%27.3max effort · Sierra4 Aug 2026Sierra Research
EBR-benchReasoning39.5%76.3max effort—Epoch AI
Blueprint-Bench 2Multimodal38.6%———Google DeepMind
Furniture AssemblyReasoning35.8%65.0max effort—Epoch AI
Terminal-Bench 3.0Agentic34.0%80.6——Ryan Marten et al.
ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobAgentic34.0%79.1——NeoCognition
Critical Physics TasksReasoning28.6%95.0——Artificial Analysis
Terminal-Bench 4.0Agentic22.7%75.5—21 Sept 2026Vals AI
Agent Arena steerabilityAgentic10.679.3high effort15 Sept 2026LMArena
Agent Arena command recoveryAgentic9.177.6high effort15 Sept 2026LMArena
Agent Arena task outcomeAgentic6.074.1high effort15 Sept 2026LMArena
ProgramBenchCoding2.0%——21 Sept 2026Vals AI
FrontierMath-ErdosMath0.0%—max effort—Epoch AI

53 benchmarks count, from 59 of 66 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedEpoch AI, collected directlyCC BY — free to use and redistribute with attributionARC Prize Foundation, collected directlyNo licence stated. Their terms ask for written permission before commercial useLiveBench, collected directlyNo licence stated for the leaderboard. The site repo serving these CSVs has no LICENSE; the harness repo carries upstream Apache-2.0 and MIT copies that cover code, not resultsVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AISierra Research, collected directlyMIT — results are in the licensed repositoryLMArena, collected directlyCC BY 4.0 (lmarena-ai/leaderboard-dataset on Hugging Face)

Same level, lower price

Muse Spark 1.375.0 · $4.25GPT-5.6 Sol76.7 · $10GPT-6 Sol76.9 · $10

More from Anthropic

Claude Opus 5.582.8Claude Fable 5.181.4Claude Opus 577.9Claude Opus 4.871.1Claude Sonnet 567.6