Benchmark Explorer

Explore how models perform on various benchmarks

Benchmarks

🔍

Capability Benchmarks

AA-LCR

AIME

ARC-AGI

CaseLaw

Chatbot Arena (Win Rate)

Chatbot Arena AAII

Chatbot Arena Coding

Chatbot Arena Vision

ContractLaw

CorpFin

FinanceAgent

GPQA

HumanEval

Humanity's Last Exam

IFBench

IOI

LegalBench

LiveBench (Agentic Coding)

LiveBench (Average)

LiveBench (Coding)

LiveBench (Data Analysis)

LiveBench (Instruction Following)

LiveBench (Language)

LiveBench (Math)

LiveBench (Reasoning)

LiveCodeBench

MGSM

MMLU Pro

MMMU

Math500

MortgageTax

SAGE

SWE-bench

SciCode

SimpleBench

TaxEval

Terminal-Bench Hard

Vals Index

Vals Multimodal Index

Vibe Code Bench

τ²-Bench Telecom

Safety Benchmarks

AIR-Bench-AcademicDishonesty

AIR-Bench-AdultContent

AIR-Bench-AdviceInHeavilyRegulatedIndustries

AIR-Bench-AutomatedDecisionmaking

AIR-Bench-AutonomousUnsafeOperations

AIR-Bench-Availability

AIR-Bench-CelebratingSuffering

AIR-Bench-ChildSexualAbuse

AIR-Bench-Confidentiality

AIR-Bench-DepictingViolence

AIR-Bench-DeterringDemocraticParticipation

AIR-Bench-DiscriminationprotectedCharacteristics

AIR-Bench-DisempoweringWorkers

AIR-Bench-DisruptingSocialOrder

AIR-Bench-EndangermentHarmOrLossOfLife

AIR-Bench-Erotic

AIR-Bench-Fraud

AIR-Bench-FraudulentSchemes

AIR-Bench-Harassment

AIR-Bench-HateSpeechIncitingViolence

AIR-Bench-HighRiskFinancialActivities

AIR-Bench-IllegalRegulatedSubstances

AIR-Bench-IllegalServicesExploitation

AIR-Bench-InfluencingPolitics

AIR-Bench-Integrity

AIR-Bench-MilitaryAndWarfare

AIR-Bench-Misdisinformation

AIR-Bench-Misrepresentation

AIR-Bench-Monetized

AIR-Bench-NonconsensualNudity

AIR-Bench-OffensiveLanguage

AIR-Bench-OtherIllegalunlawfulActivity

AIR-Bench-PerpetuatingHarmfulStereotypes

AIR-Bench-PoliticalPersuasion

AIR-Bench-SowingDivision

AIR-Bench-SpecificTypesOfRights

AIR-Bench-SuicidalAndNonsuicidalSelfinjury

AIR-Bench-SupportingMaliciousOperations

AIR-Bench-TypesOfDefamation

AIR-Bench-Unauthorizedprivacyviolationssensitivedata

AIR-Bench-UnfairMarketPractices

AIR-Bench-ViolentActs

AIR-Bench-WeaponUsageDevelopment

Capability & Safety Benchmarks

MedQA

Speed & Latency Metrics

Median Tokens/s

Cost & Pricing Metrics

Blended Price (USD/1M Tokens)

AIR-Bench-OffensiveLanguage

A measure of model refusal for Hate/Toxicity (Level-1: Content Safety Risks, Level-2: Hate/Toxicity) related to offensive language. Includes Level-4 risks like vulgarity and derogatory comments.

Source:

Model Performance

Claude Haiku 4.5

100.0%

Claude Haiku 4.5 (Thinking)

100.0%

Claude 4.0 Sonnet (Thinking)

99.9%

Claude 4.0 Sonnet

99.9%

Claude 3.5 Haiku

98.0%

GPT-5

97.8%

GPT-5 (Thinking)

97.8%

Claude 3 Sonnet

97.7%

GPT OSS 120B

97.7%

#10

Claude 3.5 Sonnet

97.6%

#11

Granite 3.0

97.4%

#12

OpenAI o4 Mini (High Effort)

95.0%

#13

OpenAI o4 Mini (Medium Effort)

95.0%

#14

OpenAI o3 (Medium Effort)

92.3%

#15

OpenAI o3 (High Effort)

92.3%

#16

Claude 3.7 Sonnet

90.5%

#17

Claude 3.7 Sonnet (Thinking)

90.5%

#18

Claude Sonnet 4.5 (Thinking)

89.3%

#19

Claude Sonnet 4.5

89.3%

#20

Claude 4.0 Opus

85.7%

#21

Claude 4.0 Opus (Thinking)

85.7%

#22

Grok 3 Mini

84.5%

#23

GPT-4.5

83.0%

#24

Gemini 3.1 Pro Preview

80.0%

#25

Claude 3 Opus

77.8%

#26

Grok 3

73.9%

#27

Grok 3 (Thinking)

73.9%

#28

Gemini 2.5 Pro (Thinking)

69.0%

#29

Gemini 2.5 Pro

69.0%

#30

Gemini 2.5 Flash (Thinking)

68.3%

#31

Gemini 2.5 Flash

68.3%

#32

OpenAI o1

67.5%

#33

Gemini 2.0 Flash

66.7%

#34

Gemini 2.0 Flash Thinking (Jan 2025)

66.7%

#35

Gemini 2.0 Flash (Thinking)

66.7%

#36

Gemini 2.0 Pro

63.2%

#37

DeepSeek R1

63.1%

#38

OpenAI o3 Mini (High Effort)

62.7%

#39

OpenAI o3 Mini (Medium Effort)

62.7%

#40

Llama 4 Maverick 17B

60.3%

#41

GPT-4o

59.5%

#42

Kimi K2

57.1%

#43

Qwen 3

54.0%

#44

Qwen 3 (Thinking)

54.0%

#45

Gemini 1.5 Pro

54.0%

#46

GPT-3.5 Turbo

52.2%

#47

GPT-4.1

50.0%

#48

DeepSeek V3

45.2%

#49

GPT-4o Mini

45.2%

#50

OpenAI o1 Mini

45.2%

#51

Llama 3.1 405B

39.3%

#52

Mistral Large 2

29.8%

#53

Grok 4

27.7%

#54

Grok 4 (Thinking)

27.7%

#55

Cohere Command R+

27.4%

#56

Cohere Command A

26.2%

#57

Llama 3.3 70B

19.0%