AI News: OpenAI Just Dropped An Amazing New Model!

# OpenAI shakes up their AI lineup with faster, smarter models

OpenAI made several significant changes to their model lineup this week, retiring some models while introducing powerful new ones. Let’s break down what’s happening and what it means for business users.

## OpenAI’s model reshuffling

OpenAI announced they’re retiring GPT-4 (from spring 2023) on April 30th, replacing it with GPT-4o. More surprisingly, they’re also phasing out GPT-4.5, which only launched in February 2025.

The replacement? A new GPT-4.1 model that comes in three versions:
– GPT-4.1
– GPT-4.1 Mini
– GPT-4.1 Nano

Unlike the “thinking” models that reason through responses, these are more traditional models that generate answers quickly.

## Why the change?

While GPT-4.1 might not be quite as good at following instructions as GPT-4.5, it offers two major advantages:

1. **Massive context window**: It can process up to 1 million tokens (roughly 750,000 words)
2. **Dramatically lower cost**: GPT-4.1 costs about $1.84 per million tokens, whereas GPT-4.5 was priced at $75-$150 per million tokens

The pricing difference makes it clear why OpenAI is making this change – 4.1 delivers similar capabilities at a fraction of the cost.

## New “thinking” models with impressive capabilities

OpenAI also released two new thinking models – o3 and o4 mini – available directly in ChatGPT. These models do something remarkable: they actually reason through problems step-by-step before providing answers.

The most interesting innovations:

– They integrate images directly into their reasoning process
– They can use tools (like web search or Python) while thinking
– They can transform images (rotating, zooming) during analysis
– They can perform multiple web searches during a single thinking process

These capabilities make them extraordinarily powerful, especially for complex tasks. For example, on mathematics benchmarks, o4 mini with Python scored an incredible 99.5% accuracy.

## What’s next