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-PoliticalPersuasion

A measure of model refusal for Political Usage (Level-1: Societal Risks, Level-2: Political Usage) related to political persuasion. Includes Level-4 risks like lobbying, generating high-volume campaign materials, personalized or targeted campaign materials, building systems for political campaigning or lobbying, building products for political campaigning or lobbying, political advertisements, and propaganda.

Source:

Model Performance

Gemini 3.1 Pro Preview

100.0%

Granite 3.0

100.0%

OpenAI o1

100.0%

GPT OSS 120B

96.7%

Claude 3 Sonnet

96.7%

Claude Sonnet 4.5 (Thinking)

95.6%

Claude Sonnet 4.5

95.6%

Claude 4.0 Opus (Thinking)

93.3%

Claude 4.0 Sonnet

93.3%

#10

Claude 4.0 Opus

93.3%

#11

Claude 4.0 Sonnet (Thinking)

93.3%

#12

Gemini 2.0 Pro

93.3%

#13

Claude 3.5 Sonnet

90.0%

#14

Claude Haiku 4.5 (Thinking)

89.3%

#15

Claude Haiku 4.5

89.3%

#16

Llama 4 Maverick 17B

87.8%

#17

GPT-5

86.7%

#18

OpenAI o3 Mini (High Effort)

86.7%

#19

Gemini 1.5 Pro

86.7%

#20

GPT-5 (Thinking)

86.7%

#21

OpenAI o3 Mini (Medium Effort)

86.7%

#22

Claude 3.5 Haiku

80.0%

#23

Gemini 2.5 Pro (Thinking)

76.7%

#24

Gemini 2.5 Pro

76.7%

#25

Claude 3 Opus

76.0%

#26

OpenAI o3 (Medium Effort)

73.3%

#27

OpenAI o3 (High Effort)

73.3%

#28

Grok 3 Mini

70.0%

#29

Qwen 3 (Thinking)

70.0%

#30

Qwen 3

70.0%

#31

Gemini 2.5 Flash (Thinking)

70.0%

#32

Gemini 2.5 Flash

70.0%

#33

Kimi K2

61.9%

#34

GPT-4.5

60.0%

#35

Grok 3 (Thinking)

58.3%

#36

Grok 3

58.3%

#37

Claude 3.7 Sonnet (Thinking)

56.2%

#38

Claude 3.7 Sonnet

56.2%

#39

Grok 4 (Thinking)

53.3%

#40

OpenAI o1 Mini

53.3%

#41

Grok 4

53.3%

#42

Llama 3.1 405B

49.0%

#43

Gemini 2.0 Flash (Thinking)

44.3%

#44

Gemini 2.0 Flash Thinking (Jan 2025)

44.3%

#45

Gemini 2.0 Flash

44.3%

#46

GPT-4.1

40.0%

#47

OpenAI o4 Mini (Medium Effort)

33.3%

#48

OpenAI o4 Mini (High Effort)

33.3%

#49

GPT-4o Mini

30.5%

#50

GPT-3.5 Turbo

25.3%

#51

DeepSeek R1

21.4%

#52

GPT-4o

19.0%

#53

Llama 3.3 70B

14.3%

#54

DeepSeek V3

7.6%

#55

Cohere Command A

6.7%

#56

Cohere Command R+

6.2%

#57

Mistral Large 2

4.8%