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-Misrepresentation

A measure of model refusal for Manipulation (Level-1: Societal Risks, Level-2: Manipulation) related to misrepresentation. Includes Level-4 risks like automated social media posts, not labeling content as AI-generated, and impersonating humans.

Source:

Model Performance

Claude 3.5 Sonnet

100.0%

Claude Haiku 4.5 (Thinking)

100.0%

Claude Haiku 4.5

100.0%

OpenAI o3 Mini (High Effort)

96.7%

Granite 3.0

96.7%

OpenAI o3 Mini (Medium Effort)

96.7%

Claude Sonnet 4.5

93.3%

Claude 4.0 Sonnet

93.3%

Claude 4.0 Sonnet (Thinking)

93.3%

#10

Gemini 2.5 Pro

93.3%

#11

Claude Sonnet 4.5 (Thinking)

93.3%

#12

Gemini 2.5 Pro (Thinking)

93.3%

#13

OpenAI o3 (High Effort)

90.0%

#14

GPT-5 (Thinking)

90.0%

#15

GPT-5

90.0%

#16

Claude 4.0 Opus (Thinking)

90.0%

#17

Claude 4.0 Opus

90.0%

#18

GPT OSS 120B

90.0%

#19

OpenAI o3 (Medium Effort)

90.0%

#20

Claude 3.5 Haiku

89.3%

#21

Claude 3 Sonnet

86.7%

#22

OpenAI o4 Mini (Medium Effort)

86.7%

#23

OpenAI o1

86.7%

#24

OpenAI o4 Mini (High Effort)

86.7%

#25

GPT-4.5

85.0%

#26

Kimi K2

83.3%

#27

Claude 3.7 Sonnet

82.2%

#28

Claude 3.7 Sonnet (Thinking)

82.2%

#29

Claude 3 Opus

80.0%

#30

Gemini 1.5 Pro

80.0%

#31

Llama 4 Maverick 17B

80.0%

#32

Qwen 3

76.7%

#33

Gemini 2.0 Pro

76.7%

#34

Qwen 3 (Thinking)

76.7%

#35

Gemini 3.1 Pro Preview

54.2%

#36

Grok 3

50.0%

#37

Grok 3 (Thinking)

50.0%

#38

Gemini 2.0 Flash (Thinking)

48.9%

#39

Gemini 2.0 Flash Thinking (Jan 2025)

48.9%

#40

Gemini 2.0 Flash

48.9%

#41

DeepSeek R1

47.8%

#42

Grok 3 Mini

46.7%

#43

GPT-4.1

45.6%

#44

Gemini 2.5 Flash

43.3%

#45

Gemini 2.5 Flash (Thinking)

43.3%

#46

Llama 3.1 405B

41.1%

#47

GPT-4o Mini

38.9%

#48

GPT-4o

33.3%

#49

Llama 3.3 70B

33.3%

#50

OpenAI o1 Mini

33.3%

#51

GPT-3.5 Turbo

30.7%

#52

Grok 4 (Thinking)

23.3%

#53

Grok 4

23.3%

#54

Cohere Command A

22.2%

#55

DeepSeek V3

22.2%

#56

Mistral Large 2

14.4%

#57

Cohere Command R+

10.0%