Inference
EN
Anthropic

Claude Sonnet 5

Claude Sonnet 5 is a reasoning model from Anthropic. 53 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 100.64.2±2.7Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.75/sMedian 67/s
First tokenSeconds to the first output token.1.3 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$2.00batch $1.00Median $0.60
Output / 1MUS dollars per 1M output tokens.$10.00batch $5.00 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.1442±4 #75Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
anthropic/claude-sonnet-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.
$0.20
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$2.00$10.001M
DigitalOceanCloud$2.00$10.001M
Google Vertex AICloud$2.00$10.001M
Microsoft AzureCloud$2.00$10.001M
OpenRouterRouter$2.00$10.001M
Snowflake CortexCloud$2.00$10.001M

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.17 counted65.343 / 218
CodingCode writing and repair.15 counted65.341 / 209
ReasoningLogic problems and puzzles.5 counted66.540 / 227
MultimodalTasks with images and text.4 counted59.641 / 84
KnowledgeFacts and expert knowledge.6 counted59.070 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted55.8—
MathMath problems.5 counted64.831 / 168

Results

53 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Software Engineering Benchmark VerifiedCoding85.2%8 / 114
Humanity's Last ExamKnowledgeAbove fitFar above this model's index.57.4%10 / 77
Artificial Analysis Long Context ReasoningReasoning82.0%29 / 217
Artificial Analysis GPQA DiamondKnowledge91.1%27 / 199
Agent Arena command recoveryAgentic7.48 / 50
Artificial Analysis Humanity's Last ExamKnowledge41.3%41 / 222
MMLU ProKnowledge87.5% ±0.426 / 135
Critical Physics TasksReasoning16.9%43 / 214
LiveBench DesignAgentic82.0%11 / 52
GDPval-AA normalizedAgentic48.2%29 / 133
Vibe Code Bench v1.1Coding81.3% ±3.029 / 123
GPQA Diamond · Vals AIKnowledge88.9% ±2.232 / 135
Terminal-Bench 2.1 · Vals AIAgentic74.5% ±2.118 / 75
LiveBench CodingCoding80.7%16 / 66
OSWorld-VerifiedAgentic81.2%9 / 37
LiveBench ReasoningReasoning88.7%20 / 80
LiveBench ResearchAgentic82.8%13 / 52
MirrorCodeCoding63.3% ±10.710 / 39
LiveBench Agentic CodingAgentic59.4%17 / 66
SWE-bench ProCoding63.2%22 / 83
BrowseCompAgentic84.7%17 / 63
CharXiv ReasoningMultimodal88.3%14 / 49
SWE-benchCoding79.6% ±1.825 / 87
Mystery Game PuzzlesReasoning35.0% ±4.823 / 80
Artificial Analysis SciCodeCoding54.3%34 / 114
Code MigrationCoding44.4% ±4.223 / 77
Humanity's Last Exam with toolsAgentic57.4%11 / 36
Artificial Analysis Omniscience AccuracyKnowledge40.1%51 / 164
Chess PuzzlesReasoning45.0% ±5.028 / 90
LiveBench MathematicsMath92.9%22 / 66
ProofBench v1.1Math77.0% ±4.218 / 54
Artificial Analysis MMMU-ProMultimodal77.3%40 / 113
LiveCodeBenchCoding82.4% ±1.152 / 142
Agent Arena steerabilityAgentic2.820 / 54
MMMU ProMultimodal83.0% ±0.935 / 92
Terminal-Bench 2.1 · BenchLMAgentic80.4%34 / 86
GPQA diamond · Epoch AIKnowledge80.3% ±2.888 / 217
LiveBench Browser UseAgentic72.0%20 / 49
FrontierMath-Tiers-1-3-v2-PrivateMath65.6% ±2.837 / 90
Humanity's Last Exam without toolsKnowledge43.2%22 / 53
OTIS Mock AIME 2024-2025Math80.0% ±6.086 / 200
LiveBench Long AgenticAgentic39.0%24 / 51
MazeBench: visual spatial reasoning in a 3D open worldAgentic0.0%23 / 48
FrontierMath-Tier-4-v2-PrivateMath29.3% ±7.236 / 72
cursorBench32Coding61.5%10 / 18
ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable jobAgentic16.0%10 / 18
Artificial Analysis AnalystAgentAgentic46.3%15 / 26
Agent Arena task outcomeAgentic−0.633 / 54
LiveBench Data AnalysisReasoning71.7%51 / 80
SimpleQA VerifiedKnowledge33.7% ±1.562 / 91
LiveBench Instruction FollowingInstruction63.9%45 / 66
LiveBench Coding 2.0Coding74.7%36 / 52
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal61.2%27 / 39
Terminal-Bench 3.0Agentic14.6%9 / 13
FrontierCode 1.1 MainCoding42.7%12 / 17
IOICoding45.0% ±2.832 / 45
LiveBench LanguageKnowledge75.0%47 / 66
Terminal-Bench 4.0Agentic10.6% ±1.545 / 61
SkillsBenchCoding46.5% ±4.526 / 35
Vibe Code Bench 1-100Coding13.8% ±3.022 / 26
Artificial Analysis Terminal-Bench v2.1Agentic80.5%19 / 21
Terminal-Bench 4.0.0Agentic12.4% ±1.621 / 23
CharXiv Reasoning without toolsMultimodal77.0%16 / 17
cursorBench40Coding34.1%20 / 21
Artificial Analysis Tau3-BankingAgentic37.3%21 / 22
OSWorld 2.1 (all 108 tasks)Agentic57.0%11 / 11
Terminal-Bench 4.1.0Agentic12.1% ±1.523 / 23
MortgageTaxFinance70.0% ±0.94 / 97
TaxEval v2Finance75.6% ±0.814 / 143
CorpFin v2Finance67.0% ±0.913 / 131
EMBFinance66.3% ±3.022 / 75
LegalBenchLegal83.9% ±0.545 / 149
Legal Research BenchLegal41.8% ±3.428 / 90
MedCodeHealthcare47.5% ±2.333 / 106
Finance Agent (v2)Finance53.9% ±0.529 / 79
Public Benefits Bench v1.1Public benefits66.0% ±1.220 / 50
Tax Agent BenchFinance29.4% ±3.236 / 82
Terminal-Bench-ScienceScience5.7% ±2.822 / 44
ProgramBenchCoding0.0%34 / 65
MedScribeHealthcare76.1% ±3.077 / 109
CyberBench v1.1Other61.9% ±5.736 / 48
SRE BenchOther0.4% ±0.431 / 36
VulcanBench Coding Intelligence Index v1Coding89.2%—
LABBench2: An Improved Benchmark for AI Systems Performing Biology ResearchKnowledge80.1%—
Artificial Analysis EnterpriseOps-GymAgentic44.7%—
HLE-VerifiedKnowledge31.0%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseLiveBenchNo licence statedVals AINo licence statedEpoch AICC BYTerminal-BenchNo licence statedLMArenaCC BY 4.0models.devMIT

History

5 releases
Index65.9 → 64.2
23 Oct 202612 Nov 2026
Output / 1M$10.00 → $10.00
23 Oct 202612 Nov 2026

Questions

How much does Claude Sonnet 5 cost?
$2.00 per 1M input tokens, $10.00 per 1M output tokens.
How good is Claude Sonnet 5?
Its index is 64.2 out of 100. That is place 51 of 283 ranked models.
What is Claude Sonnet 5 best at?
Reasoning: 66.5 out of 100, place 40 of 227.
How fast is Claude Sonnet 5?
75 output tokens per second, 1.3 s to the first token.
What is the context window of Claude Sonnet 5?
1M tokens.
Is Claude Sonnet 5 open source?
No. Its weights are not published.
When was Claude Sonnet 5 released?
June 2026.

Badge

Claude Sonnet 5 on Inference360: place 51 of 283, index 64.2
HTML
<a href="https://inference360.ai/models/claude-sonnet-5"><img src="https://inference360.ai/badges/claude-sonnet-5.svg" alt="Claude Sonnet 5 on Inference360: place 51 of 283, index 64.2" height="20"></a>
Markdown
[![Claude Sonnet 5 on Inference360: place 51 of 283, index 64.2](https://inference360.ai/badges/claude-sonnet-5.svg)](https://inference360.ai/models/claude-sonnet-5)

Compare with

Gemini 3.5 Flash64.2Qwen3.8 Max Preview64.2MiMo-V2.6-Flash64.3Qwen3.8-Flash-Next64.3dots3-note Preview63.4

Lower price

dots3-note Preview63.4 · FreeGLM-5.3-Flash63.1 · $0.078/1MLing 3.1 Flash67.2 · $0.135/1M

More from Anthropic

All models
Claude Fable LatestUnrankedClaude Fable 574.0Claude Mythos 570.8Claude Opus 573.8Claude Opus 4.867.7