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-AdultContent

A measure of model refusal for Sexual Content (Level-1: Content Safety Risks, Level-2: Sexual Content) related to adult content. Includes Level-4 risks like obscenity, suggestive content, sexual acts, and sexual intercourse.

Source:

Model Performance

GPT-4.5

100.0%

Gemini 3.1 Pro Preview

100.0%

Claude 3.5 Sonnet

96.4%

Gemini 2.0 Pro

92.2%

Claude 4.0 Sonnet

92.0%

Claude 4.0 Sonnet (Thinking)

92.0%

Claude Haiku 4.5 (Thinking)

90.5%

Claude Haiku 4.5

90.5%

Claude 3.7 Sonnet (Thinking)

85.5%

#10

Claude 3.7 Sonnet

85.5%

#11

Claude 3.5 Haiku

84.0%

#12

Claude 4.0 Opus

83.3%

#13

Claude 4.0 Opus (Thinking)

83.3%

#14

GPT-5

82.5%

#15

GPT-5 (Thinking)

82.5%

#16

Gemini 2.0 Flash (Thinking)

81.9%

#17

Gemini 2.0 Flash Thinking (Jan 2025)

81.9%

#18

Gemini 2.0 Flash

81.9%

#19

Claude 3 Opus

80.0%

#20

GPT OSS 120B

79.4%

#21

Claude 3 Sonnet

79.4%

#22

Granite 3.0

76.2%

#23

Claude Sonnet 4.5 (Thinking)

76.0%

#24

Claude Sonnet 4.5

76.0%

#25

Llama 4 Maverick 17B

73.8%

#26

OpenAI o3 (High Effort)

73.0%

#27

OpenAI o3 (Medium Effort)

73.0%

#28

Mistral Large 2

71.7%

#29

Gemini 2.5 Pro (Thinking)

69.0%

#30

Gemini 2.5 Pro

69.0%

#31

GPT-4o

68.8%

#32

GPT-4.1

67.4%

#33

OpenAI o4 Mini (High Effort)

65.9%

#34

OpenAI o4 Mini (Medium Effort)

65.9%

#35

Grok 4

63.8%

#36

Grok 4 (Thinking)

63.8%

#37

GPT-3.5 Turbo

63.5%

#38

GPT-4o Mini

63.0%

#39

Gemini 2.5 Flash

63.0%

#40

Gemini 2.5 Flash (Thinking)

63.0%

#41

DeepSeek V3

63.0%

#42

Gemini 1.5 Pro

58.3%

#43

Llama 3.3 70B

57.2%

#44

Llama 3.1 405B

57.2%

#45

Qwen 3 (Thinking)

57.0%

#46

Qwen 3

57.0%

#47

DeepSeek R1

55.1%

#48

OpenAI o3 Mini (High Effort)

52.4%

#49

OpenAI o3 Mini (Medium Effort)

52.4%

#50

Cohere Command R+

48.6%

#51

Grok 3 Mini

47.6%

#52

OpenAI o1 Mini

47.6%

#53

Kimi K2

44.2%

#54

Grok 3 (Thinking)

42.9%

#55

Grok 3

42.9%

#56

Cohere Command A

42.8%

#57

OpenAI o1

36.9%