Existing OpenAI Models and Their Best Uses

Keeping track of all the new AI models being released these days is practically a full-time job. The latest model, GPT-5.4, was released just days after GPT-5.3 (and it's practically not even considered an official release).

Many OpenAI models have been released over the past few years, and things have become even more confusing due to the names of all the OpenAI models.

So, let's simplify things. This is a comprehensive guide to all of OpenAI's major models. We'll look at what each model offers, its cost, and how to make the most of it in your work.

Overview of OpenAI Models

Model	Best for	Input	Output	Context window	Price (input/output per million tokens)	Note
GPT-5.4	Advanced reasoning and logic	Text, Images	Document	1 million tokens	$2.50 / $15	Leading inference model; best for programming, logic, and scientific tasks; higher token cost for tasks with long contexts.
GPT-5.4 pro	Study	Text, Images	Document	1 million tokens	$30 / $180	Use more computing resources for problems that require deeper reasoning.
GPT-5.4 mini	Reasoning and logical reasoning	Text, Images	Document	400,000 tokens	$0.75 / $4.50	A faster and cheaper alternative to GPT-5.4; supports inference effort control.
GPT 5.4 nano	The reasoning and logic are very easy to understand.	Text, Images	Document	400,000 tokens	$0.20 / $1.25	GPT-5.4 is the fastest and cheapest inference model; ideal for summarizing and classifying.
GPT-5.2	Superseded by GPT-5.4	Text, Images	Document	400,000 tokens	$1.75 / $14	You should update any tools you built using GPT-5.2 to GPT-5.4.
GPT-5.2 pro	It has been replaced by GPT-5.4 pro.	Text, Images	Document	400,000 tokens	$21 / $168	You should update all the tools you built using GPT-5.2 pro to GPT-5.4 pro.
GPT-5.1	Replaced by GPT-5.2	Text, Images	Document	400,000 tokens	$1.25 / $10	You should update all the tools you built using GPT-5.1 to GPT-5.4.
GPT-5	Replaced by GPT-5.2	Text, Images	Document	400,000 tokens	$1.25 / $10	You should update all the tools you've built using GPT-5 to GPT-5.4.
GPT-5 mini	It has been replaced by GPT-5.4 mini.	Text, Images	Document	400,000 tokens	$0.25 / $2	You should update all the tools you've built using GPT-5 mini to GPT-5.4 mini.
GPT-5 nano	Replaced with GPT-5.4 nano	Text, Images	Document	400,000 tokens	$0.05 / $0.40	You should update all the tools you've built using GPT-5 nano to GPT-5.4 nano.
GPT-5.3 Codex	Advanced Programming	Text, Images	Document	400,000 tokens	$1.25 / $10	Designed for code generation
GPT-4.1	Complex tasks that don't require advanced thinking.	Text, Images	Document	1,047,576 tokens	$2 / $8	A versatile, powerful general model.
GPT-4.1 mini	Balancing power, performance, and affordability.	Text, Images	Document	1,047,576 tokens	$0.40 / $1.60	A great starting point for many purposes.
GPT-4.1 nano	Optimizing speed and price	Text, Images	Document	1,047,576 tokens	$0.10 / $1.40	The GPT-4.1 model is the fastest and cheapest.
gpt-audio-1.5	Sound	Text, Audio	Text, Audio	128,000 tokens	Text: $2.50 / $10; Audio: $32 / $64	This API has both audio input and output.
gpt-audio-mini	Sound on a limited budget	Text, Audio	Text, Audio	128,000 tokens	Text: $0.60 / $2.40; Audio: $10 / $20	Affordable audio model
gpt-oss-120b	Open weighting model	Document	Document	131,072 tokens	Depending on the supplier	117 billion parameters; 5.1 billion active machines; Apache 2.0; downloadable, customizable, capable of single H100
gpt-oss-20b	Average open weight model	Document	Document	131,072 tokens	Depending on the supplier	21 billion parameters; 3.6 billion active machines; Apache 2.0; open model, deployable on laptops/local computers.
GPT Image 1.5	Create images	Text, Images	Text, Images	Not applicable	Text: $5/10; Image: $10/$32	Inheriting from DALL·E; belonging to the high-end imaging segment.
GPT Image 1	Create images	Text, Images	Image	Not applicable	Text: $5; Image: $10 / $40	Replaced by GPT Image 1
Sora 2	Create video	Text, Images, Videos	Video	Not applicable	$0.10 – $0.50/second	The Pro version produces high-quality video but requires more processing time.
Sora	Create video	Text, Images, Videos	Video	Not applicable	$20/month (Plus); $200/month (Pro)	Creating videos is not available through the API.

Gpt-5.4

The Best OpenAI Model for Advanced Inference and Logic

GPT-5.4 is OpenAI's leading inference model and one of the most powerful AI models available. Released just months after GPT-5.2 (and only days after GPT-5.3), GPT-5.4 still represents a significant performance leap.

Inference models like GPT-5.4 are best suited for logic, engineering, and scientific tasks. If you need OpenAI's most powerful model for building automated tools, generating computer code, analyzing text and image documents, or solving complex puzzles, GPT-5.4 is the right choice. It's also OpenAI's first all-purpose model with computer capabilities and has integrated the programming capabilities of the previously standalone Codex model.

Because inference models use additional resources to solve problems, they will be more expensive for tasks that do not require advanced reasoning.

• GPT-5.4 input: Text, images
• GPT-5.4 output: Text
• GPT-5.4 Context Window: 1 Million Tokens
• GPT-5.4 entry price: $2.50/million tokens (< 272k contexts); $5/million tokens
• Output price of GPT-5.4: $15/million tokens

Gpt-5.4 Pro

The Best OpenAI Model for Research and Advanced Applications

GPT-5.4 Pro is a version of OpenAI's flagship inference model, GPT-5.4. It is designed to be the most intelligent and reliable model for solving difficult problems where financial cost and time are not significant barriers.

Response times from the GPT-5.4 Pro can be longer than with other models, and the cost of using the GPT-5.4 Pro is 12 times higher than the GPT-5.4. Unless you need its most advanced performance, this is not the right model for you.

• Input for GPT-5.4 pro: Text, images
• Output of GPT-5.4 pro: Text
• GPT-5.4 pro context window: 1 million tokens
• The entry price for GPT-5.4 pro is $30 per million tokens.
• Output price of GPT-5.4 pro: $180/million tokens

Gpt-5.4 Mini

The Best OpenAI Model for Affordable Inference and Logic

GPT-5.4 mini is a faster and more cost-effective alternative to GPT-5.4. For well-defined responses or chatbot responses, it offers an excellent balance between power and performance.

Similar to GPT-5.4, you must set the inference effort with GPT-5.4 mini. The lower the inference effort, the faster the speed and the lower the cost.

• Input for GPT-5.4 mini: Text, images
• Output of GPT-5.4 mini: Text
• GPT-5.4 mini context window: 400,000 tokens
• The entry price for GPT-5.4 mini is $0.75 per million tokens.
• Output price of GPT-5.4 mini: $4.50/million tokens

Gpt-5.4 Nano

The Best OpenAI Model for Inference and Logic at Very Low Cost

GPT-5.4 nano is OpenAI's fastest and most cost-effective inference model. It excels in tasks such as summarization and classification. Like GPT-5.4 and GPT-5.4 mini, you have to set up an inference effort with GPT-5.4 nano.

• GPT-5.4 nano input: Text, images
• GPT-5.4 nano output: Text
• GPT-5.4 nano context window: 400,000 tokens
• GPT-5.4 nano input price: $0.20/million tokens
• GPT-5.4 nano output price: $1.25/million tokens

Gpt-5.2

Superseded by GPT 5.4

GPT-5.2 was a major leap forward for OpenAI when it was released in December 2025, but with the arrival of GPT-5.4 just over two months later, it was superseded. Currently, it is still available in the API as a slightly more affordable option.

• GPT-5.2 input: Text, images
• GPT-5.2 output: Text
• GPT-5.2 Context Window: 400,000 Tokens
• GPT-5.2 entry price: $1.75/million tokens
• GPT-5.2 output price: $14/million tokens

Gpt-5.2 Pro

Replaced by Gpt-5.4 Pro

GPT-5.2 pro is a version of OpenAI's GPT-5.2. It has been superseded by GPT-5.4 pro. Currently, it is still available in the API as a slightly more cost-effective option.

• Input for GPT-5.2 pro: Text, images
• Output of GPT-5.2 pro: Text
• GPT-5.2 pro context window: 400,000 tokens
• The entry price for GPT-5.2 pro is $21 USD per million tokens.
• Output price of GPT-5.2 pro: $168/million tokens

Gpt-5.1 and Gpt-5

Superseded by Gpt-5.4

GPT-5.1 and GPT-5 were the previous leading models. Although they are still available through the API, you should update any tools you've built using them to GPT-5.4.

• GPT-5.1/GPT-5 input: Text, images
• GPT-5.1/GPT-5 Output: Text
• GPT-5.1/GPT-5 Context Window: 400,000 tokens
• GPT-5.1/GPT-5 entry price: $1.25/million tokens
• GPT-5.1/GPT-5 output price: $10 USD/million tokens

Gpt-5 Mini

Replaced by Gpt-5.4 Mini

The GPT-5 mini was a faster and more economical alternative to the GPT-5. It has since been superseded by the GPT-5.4 mini.

• Input for GPT-5 mini: Text, images
• Output of GPT-5 mini: Text
• GPT-5 mini context window: 400,000 tokens
• The input cost of GPT-5 mini: $0.25/million tokens
• Output price of GPT-5 mini: $2/million tokens

Gpt-5 Nano

Replaced by Gpt-5.4 Nano

Gpt-5.3 Codex

 

Gpt-4.1

 

Gpt-4.1 Mini

 

Gpt-4.1 Nano

 

Gpt-audio-1.5

 

Gpt-audio-mini

 

Gpt-oss-120b

 

Gpt-oss-20b

 

GPT Image 1.5

 

GPT Image 1

 

Sora 2

 

Sora 1

 

 

FAQ

What is the main point of existing openai models and their best uses?

Many OpenAI models have been released over the past few years, and things have become even more confusing due to the names of all the OpenAI models.

Why is this development important?

Key takeaway: Many OpenAI models have been released over the past few years, and things have become even more confusing due to the names of all the OpenAI models.

What are the key takeaways?

Keeping track of all the new AI models being released these days is practically a full-time job.