OpenAI announces GPT-4.5

### Summary of "OpenAI announces GPT-4.5, warns it’s not a frontier AI model" | The Verge

OpenAI has officially launched GPT-4.5, its latest and largest AI language model, which is currently available as a research preview for ChatGPT Pro users. OpenAI claims that GPT-4.5 is the "most knowledgeable model yet," but it emphasizes that this model is not a frontier AI and may not perform as well as its previous models, o1 or o3-mini.

#### Key Points:

- **Capabilities**: GPT-4.5 features enhanced writing skills, improved world knowledge, and a more refined personality compared to earlier models. It aims to provide more natural interactions and is better at recognizing patterns, making it well-suited for writing, programming, and solving practical problems.

  

- **Limitations**: Despite improvements, OpenAI clarifies that GPT-4.5 does not introduce significant new capabilities to be classified as a frontier model. The model improves computational efficiency by more than 10x compared to GPT-4 but does not add seven new frontier capabilities. Its performance remains below that of the o1 and o3-mini models in most evaluations.

- **Training Methods**: OpenAI has used a combination of new supervision techniques together with traditional methods like supervised fine-tuning and reinforcement learning from human feedback to train GPT-4.5. 

- **Performance**: The new model reportedly hallucinates less than GPT-4o and slightly less than the o1 model. OpenAI researchers reported that GPT-4.5 performs better in various categories when compared to GPT-4o during user testing.

- **Release Schedule**: After launching for Pro users, GPT-4.5 will roll out to Plus and Team users next week, followed by Enterprise and Edu users. It is also available on Microsoft’s Azure AI Foundry platform.

- **Future Models**: OpenAI plans to release GPT-5, which is expected to integrate various technologies and include the new o3 reasoning model, by late May.

#### Notable Quotes:

- "GPT-4.5 is not a frontier model, but it is OpenAI’s largest LLM."

- "We aligned GPT-4.5 to be a better collaborator, making conversations feel warmer, more intuitive, and emotionally nuanced."

This announcement highlights OpenAI's continuous efforts to develop advanced AI models while managing expectations regarding their capabilities and performance in real-world applications. 

For more details, you can read the full article [here](https://www.theverge.com/news/620021/openai-gpt-4-5-orion-ai-model-release).