GPT-4.5 at a glance
GPT-4.5 scaled up unsupervised pre-training for broader knowledge and fewer hallucinations, launching first for ChatGPT Pro and as an expensive API preview.
What changed with GPT-4.5
OpenAI released GPT-4.5 on February 27, 2025, as a research preview of what it called its largest and best model for chat. Rather than reasoning step by step, it relied on scaled-up pre-training and post-training to improve knowledge, intent following and writing, and OpenAI said it expected the model to hallucinate less.
ChatGPT Pro users got it first, with Plus, Team, Enterprise and Edu following over the next two weeks. Developers on all paid tiers could preview it in the Chat Completions, Assistants and Batch APIs with function calling, Structured Outputs and image input. OpenAI warned it was very compute-intensive, more expensive than GPT-4o, and might not stay in the API long term.
Where to compare GPT-4.5
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Official resources
Why GPT-4.5 matters for businesses
For businesses, GPT-4.5 was a signal that bigger pre-training alone gave diminishing returns for the price, so most production workloads stayed on cheaper models while reasoning models took the lead for complex tasks.
Sources for this story
- Introducing GPT-4.5 (opens in a new tab)Official
OpenAI · openai.com
- GPT-4.5 (Preview) model page (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #openai
- #gpt-4-5
