Skip to content
Skip to main content
header search
NeuralForge
Menu
sticky search

OpenAI launches GPT-4.5 AI model with deeper knowledge and emotions

OpenAI has just introduced its latest AI model, dubbed GPT-4.5, which the company claims is its largest and best model yet. This is not a reasoning model, a faculty which is limited to the O-series models. Despite that, GPT-4.5 is touted to be a more natural conversationalist with a higher emotional quotient and improved problem-solving capabilities.

As far as core competencies go, it has access to the latest information from the web, allows file and multimedia upload, and supports canvas platform for coding-related tasks. However, it currently lacks support for voice mode, video comprehension, and screen-sharing.

Recommended Videos

For now, this model is limited to the research preview phase, which means it’s not even widely available to ChatGPT Plus subscribers. It seems infrastructure challenges are holding back the wider availability of GPT-4.5 to the market.

Benchmark comparison for OpenAI GPT 4.5 model
OpenAI

“It is a giant, expensive model. we really wanted to launch it to plus and pro at the same time, but we’ve been growing a lot and are out of GPUs. we will add tens of thousands of GPUs next week and roll it out to the plus tier then,” explained OpenAI CEO, Sam Altman.

Related: 
‘I lost my only friend’: GPT-5 gave ChatGPT a new personality, and it’s polarizing

Unlike models that reason through a chain of thought and then provide answers, GPT-4.5 relies on a technique called unsupervised learning, that involves scaling up the training data as well as the compute requirements.

GPT-4.5 is ready!

good news: it is the first model that feels like talking to a thoughtful person to me. i have had several moments where i've sat back in my chair and been astonished at getting actually good advice from an AI.

bad news: it is a giant, expensive model. we…

— Sam Altman (@sama) February 27, 2025

The overall benefit is reduced hallucination (one of the biggest trust problems with any generative AI chatbot), a broader base of knowledge, and deeper understanding.

As per benchmarks shared by the company, it ranks much higher at result accuracy compared to the GPT-4o, o3-mini, and o1 AI models. At the same time, the hallucination rate is less than half compared to the o3-mini model, and lower than the rest of its siblings.

OpenAI GPT 4.5 performance analysis.
OpenAI

“For GPT‑4.5, we developed new, scalable techniques that enable training larger and more powerful models with data derived from smaller models,” adds the company. It can handle creative queries, professional questions, and nuanced conversations in a more natural way than any model that OpenAI has launched so far.

Overall, the AI giant says GPT-4.5 is better at creativity, aesthetic intelligence, writing, as well as design-related chores. Access to GPT 4.5 is now rolling out to ChatGPT Pro subscribers on web, mobile, and desktop. ChatGPT Plus and Team tiers will get access to it next week, followed by Enterprise and Education bundles.

More exciting content