OpenAI Launches Powerful New Reasoning AI Models with Bonus Features

On Wednesday, OpenAI unveiled its newest AI reasoning models, o3 and o4-mini. These new models are built to carefully analyze questions before responding, improving their ability to handle complex tasks and provide more thoughtful answers.

O3: OpenAI’s Most Advanced Reasoning AI

The O3 model takes the lead as OpenAI’s most advanced reasoning AI so far. It outperforms earlier models in various areas, including math, coding, science, and visual understanding. 

OpenAI reports that O3 has achieved a top-tier performance in coding tests, with an impressive score of 69.1% on the SWE-bench test.

O4-Mini: A Smart Trade-Off Between Price and Performance

On the other hand, the o4-mini model strikes a balance between cost, speed, and performance. It’s a great option for developers looking for efficiency without sacrificing quality. 

While it scores 68.1% on the same SWE-bench test, just shy of o3’s performance, o4-mini offers a more budget-friendly solution.

Developers now have the flexibility to choose between o3 and o4-mini based on their needs, with o4-mini offering a more affordable option while still delivering solid results.

Read More = OpenAI Plans to Watermark ChatGPT Images — But There’s a Twist

Thinking with Images: A Key Feature

Both O3 and O4-mini come with a standout feature — the ability to reason through images. Unlike previous models, these can now analyze and interpret visual content such as sketches, diagrams, and even blurry or low-quality images. 

The models can zoom in, rotate, and process these images during their reasoning process, making them much more reliable.

Code Execution and Web Search Capabilities

In addition to handling images, o3 and o4-mini also support code execution. With the new Canvas feature, users can run Python code directly within Chatgpt and interact with dynamic data in real-time. 

Plus, these models can now search the web for up-to-date information, keeping them informed about the latest events.

Pricing and Availability

These new models are available to Openai Pro, Plus, and Team subscribers. For developers, o3 is priced at $10 per million input tokens (around 750,000 words) and $40 per million output tokens. 

Meanwhile, o4-mini comes at a much lower cost, with $1.10 per million input tokens and $4.40 per million output tokens.

Along with ChatGPT, these models — o3, o4-mini, and o4-mini-high — are also available through OpenAI’s developer APIs, allowing businesses and developers to integrate these advanced features into their applications.

O3-Pro and the Future with GPT-5

Looking ahead, OpenAI plans to release o3-pro, an enhanced version of o3 that uses additional computing power to provide even more precise answers. This version will be available exclusively to ChatGPT Pro subscribers.

Moreover, OpenAI hints that o3 and o4-mini might be among the last standalone reasoning models before GPT-5 arrives. This upcoming model is expected to merge traditional models like GPT-4.1 with advanced reasoning models, paving the way for even more powerful AI systems.