GPT-4o
text image audio
paid
OpenAI's flagship multimodal model capable of reasoning across text, images, and audio in real time.
Version: 4o
Released: 1y 5m 19d ago on 05/13/2024
Updated: 9m 22d ago on 01/10/2025
Pricing:
- input_per_1k_tokens: 0.005
- output_per_1k_tokens: 0.015
- currency: USD
- subscription_available: true
Architecture
- family: GPT
- parameters: Unknown (est. 1.8T)
- training_data: Web text, code, multimodal data
- context_length: 128000
- inference_type: cloud
Capabilities
- chat
- text-generation
- image-understanding
- speech-to-text
- multimodal-reasoning
- translation
- code generation
Languages Supported
enesfrdezhjakoarhi
Benchmarks
- MMLU: 88.5
- GSM8K: 94
- HumanEval: 82
- ImageQA: 87.2
Safety
- Extensive content filtering
- RLHF and system prompt guardrails
- Moderate bias mitigation; some Western cultural preference noted
Deployment
- regions: global
- hosting: OpenAI Cloud
- integrations: Azure OpenAI Service, ChatGPT UI, API SDKs
API Access
Auth: API Key
Tags
proprietarymultimodalpaidstate-of-the-art