No longer supported

This model is legacy (released November 2023) and is no longer actively maintained or recommended by OpenAI. Consider using their current flagship model instead.

legacyfrontiermultimodal Proprietary

GPT-4-vision-preview

by OpenAI· Released November 2023· Cutoff April 2023

GPT-4-vision-preview is a multimodal variant of GPT-4 that can analyze images and provide textual responses about them. It was released as a preview model in November 2023, offering vision capabilities alongside the language understanding of GPT-4. This model is now legacy, superseded by GPT-4o and GPT-4 Turbo with Vision.

Official Site API Docs

Input cost

$10.00 per 1M tokens

Output cost

$30.00 per 1M tokens

Context window

128K tokens

Max output

4096 tokens

Modalities

textimage

License

proprietary

Capabilities

VisionImage UnderstandingText GenerationCode GenerationFunction CallingStreamingJSON Mode

Best For

Analyzing images and generating detailed textual descriptions or answers about visual content.

Strengths

Strong image understanding and reasoning
Large 128K context window
Supports function calling and structured outputs
High-quality text generation inherited from GPT-4

Limitations

Legacy model, no longer recommended for new projects
Higher cost compared to newer models like GPT-4o
Limited to image input (no video or audio)
Slower inference than optimized models

Use Cases

Image captioning and description

Visual question answering

Document analysis (e.g., invoices, forms)

Content moderation of images

Assistive technology for visually impaired

Educational tools for explaining diagrams

E-commerce product image analysis

Improvements Over Previous Model

Introduced vision capabilities to GPT-4, enabling image input and analysis
Maintained the 128K context window from GPT-4 Turbo
Supported function calling and JSON mode for structured outputs
Provided a preview for developers to test multimodal features before full release

Back to all models