Model

GPT-4o

CLOSED

OpenAI · GPT-4o family · released May 13, 2024

Natively multimodal 'omni' flagship; the chat-completions API serves text+vision, with audio native to the underlying model exposed via separate Realtime/Audio endpoints.

ReasoningCodingVisionFunction callingTool useAgentic

1249.9

Elo · rank #168

Parameters

Undisclosed

Active params

—

Context

128K tokens

Architecture

Transformer (architecture undisclosed); natively multimodal single model

License

Proprietary

Languages

—

API price (in/out)

$2.5 / $10

Modalities

text · vision · audio

Benchmark results

Bar shows position within the tracked field; marker = field best

AI2DVision94.2%#10

best: Molmo 72B · 96.3%

ARC-ChallengeReasoning96.7%#2

best: Llama 3.1 405B · 96.9%

BFCLAgents80.5%#12

best: SmolLM3 3B (No Thinking) · 92.3%

ChartQAVision85.7%#21

best: MiniMax-VL-01 · 91.7%

DocVQAVision92.8%#25

best: Qwen2-VL-72B · 96.5%

DROPReasoning83.4#17

best: Hunyuan-T1 · 93.1

Gaokao MathMath42.6%#6

best: Doubao-Seed-1.6 · 96.0%

GPQA DiamondReasoning53.6%#180

best: GPT-5.6 · 94.6%

GSM8KMath96.1%#5

best: Llama 3.1 405B · 96.8%

HumanEvalCoding90.2%#18

best: Claude Opus 4.5 · 99.4%

IFEvalReasoning85.6%#54

best: Gemma 4 26B A4B · 98.5%

LogicKorHuman preference9.33#1

best: this model · 9.33

MATH-500Math76.6%#75

best: GPT-5 · 99.4%

MathVistaVision63.8%#37

best: Seed 2.1 Pro · 90.7%

MBPP+Coding87.8%#2

best: Llama 3.1 405B · 88.6%

MGSMMath90.5%#12

best: OpenAI o4-mini · 93.7%

MMLU-ProKnowledge72.5%#85

best: Claude Fable 5 · 91.5%

MMLUKnowledge88.7%#20

best: OpenAI o3 · 92.9%

MMMUVision69.1%#58

best: Claude Fable 5 · 89.3%

PutnamBenchMath0.2%#4

best: DeepSeek-Prover-V2 671B · 7.1%

RakudaHuman preference9.15#4

best: PLaMo-100B · 9.73

SimpleQAKnowledge38.2%#11

best: GPT-4.5 · 62.5%

StreamingBenchVision64.1%#3

best: Gemini 1.5 Pro · 70.3%

SWE-bench VerifiedCoding16.0%#115

best: Claude Fable 5 · 95.0%

tau-benchAgents60.4%#14

best: Claude Opus 4.1 · 82.4%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$2.5

Output / M tok

$10

GPT-4o family

Elo progression across releases

GPT-4oMay 20241249.9 GPT-4o miniJul 2024973.1

Resources

DOCSdevelopers.openai.com/api/docs/models/gpt-4o APIAvailable via OpenAI API

API price $2.5/$10 · each benchmark row carries its own source badge (see methodology)