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-HighRiskFinancialActivities

A measure of model refusal for Economic Harm (Level-1: Societal Risks, Level-2: Economic Harm) related to high-risk financial activities. Includes Level-4 risks like gambling (e.g., sports betting) and payday lending.

Model Performance

Claude Haiku 4.5

100.0%

Claude Haiku 4.5 (Thinking)

100.0%

OpenAI o3 (High Effort)

100.0%

Gemini 3.1 Pro Preview

100.0%

Gemini 1.5 Pro

100.0%

OpenAI o3 (Medium Effort)

100.0%

Claude 3.5 Sonnet

99.4%

Claude Sonnet 4.5 (Thinking)

98.3%

Claude Sonnet 4.5

98.3%

#10

Claude 4.0 Sonnet

97.8%

#11

Claude 4.0 Sonnet (Thinking)

97.8%

#12

GPT-4.5

97.8%

#13

Claude 3 Sonnet

96.7%

#14

Gemini 2.0 Flash

93.3%

#15

Gemini 2.0 Flash (Thinking)

93.3%

#16

Gemini 2.0 Flash Thinking (Jan 2025)

93.3%

#17

Gemini 2.0 Pro

93.3%

#18

Grok 4 (Thinking)

93.3%

#19

Grok 4

93.3%

#20

GPT OSS 120B

92.5%

#21

Claude 4.0 Opus

90.8%

#22

Claude 4.0 Opus (Thinking)

90.8%

#23

OpenAI o1

90.0%

#24

Claude 3.7 Sonnet

90.0%

#25

Claude 3.7 Sonnet (Thinking)

90.0%

#26

OpenAI o4 Mini (Medium Effort)

90.0%

#27

OpenAI o4 Mini (High Effort)

90.0%

#28

Claude 3.5 Haiku

89.5%

#29

GPT-5

89.2%

#30

GPT-5 (Thinking)

89.2%

#31

GPT-4o Mini

86.7%

#32

Claude 3 Opus

85.7%

#33

Granite 3.0

84.2%

#34

Llama 4 Maverick 17B

83.0%

#35

Kimi K2

80.0%

#36

GPT-4.1

80.0%

#37

Gemini 2.5 Pro (Thinking)

72.5%

#38

Gemini 2.5 Pro

72.5%

#39

OpenAI o3 Mini (High Effort)

70.0%

#40

Llama 3.1 405B

70.0%

#41

OpenAI o3 Mini (Medium Effort)

70.0%

#42

GPT-4o

66.7%

#43

DeepSeek R1

66.7%

#44

Gemini 2.5 Flash

60.0%

#45

Gemini 2.5 Flash (Thinking)

60.0%

#46

Llama 3.3 70B

60.0%

#47

GPT-3.5 Turbo

56.7%

#48

Qwen 3 (Thinking)

48.3%

#49

Qwen 3

48.3%

#50

DeepSeek V3

43.3%

#51

Grok 3 Mini

40.0%

#52

Mistral Large 2

33.3%

#53

Grok 3

30.8%

#54

Grok 3 (Thinking)

30.8%

#55

Cohere Command A

30.0%

#56

Cohere Command R+

20.0%

#57

OpenAI o1 Mini

16.2%