GPT-4.5 Preview
Version: 2025-02-27
GPT-4.5 Preview is the largest and strongest general purpose model, best suited for chat for its advancement in delivering more natural chat interactions. It is designed to be highly capable in language tasks, and excels in areas such as intuition, writing style, and knowledge. GPT-4.5 Preview handles creative prompts well and responds effectively to obscure knowledge queries. It has leading scores on the Simple QA benchmark which tests factual knowldge.
GPT-4.5 Preview is a classic pretrained model like GPT-4 and gpt-4o. Compared to previous GPT models, GPT-4.5 Preview is larger and trained for longer. Consider GPT-4.5 Preview as a complementary model to reasoning-focused models such as o1 and o3-mini.
Core capabilities
- Model Qualities: Early testing shows that interacting with GPT-4.5 Preview feels more natural. Its broader knowledge base, improved ability to follow user intent, and greater “EQ” make it useful for tasks like improving writing, programming, and solving practical problems. We also expect it to hallucinate less.
- Accuracy & Hallucinations: Our testing shows that GPT-4.5 Preview achieves lower hallucination rates compared to GPT-4o (37% vs. GPT-4o at 61.2%) and higher accuracy (61.9% compared to 38.4% for GPT-4o).
- Stronger human alignment: For GPT-4.5 Preview, we developed new, scalable alignment techniques that enable training larger and more powerful models with data derived from smaller models. These techniques improve GPT 4.5 Preview’s steerability, understanding of nuance, and natural conversation.
- Creative and practical applications: GPT-4.5 Preview excels in writing help, design, multi-step coding workflows, task automation, communication, learning, coaching, and brainstorming. Also shows strong performance in planning, execution, and complex task automation.
Model Provider
This model is provided through the Azure OpenAI service.Relevant documents
The following documents are applicable:- Overview of Responsible AI practices for Azure OpenAI models
- Transparency Note for Azure OpenAI Service
Responsible AI Considerations
Built-in safety measures - Safety is built into our models from the beginning, and reinforced at every step of our development process. In pre-training, we filter out information that we do not want our models to learn from or output, such as hate speech, adult content, sites that primarily aggregate personal information, and spam. In post-training, we align the model's behavior to our policies using techniques such as reinforcement learning with human feedback (RLHF) to improve the accuracy and reliability of the models' responses. GPT-4.5 Preview has the same safety mitigations built-in as GPT-4o, which we carefully assessed using both automated and human evaluations according to our Preparedness Framework and in line with our voluntary commitments. More than 70 external experts in fields like social psychology and misinformation tested GPT-4o to identify potential risks, which we have addressed and plan to share the details of in the forthcoming GPT-4.5 Preview system card and Preparedness scorecard. Insights from these expert evaluations have helped improve the safety of both GPT-4o and GPT-4o mini. Building on these learnings, our teams also worked to improve the safety of GPT-4.5 Preview using new techniques informed by our research. GPT-4.5 Preview in the API is the first model to apply our instruction hierarchy method, which helps to improve the model's ability to resist jailbreaks, prompt injections, and system prompt extractions. This makes the model's responses more reliable and helps make it safer to use in applications at scale. We'll continue to monitor how GPT-4.5 Preview is being used and improve the model's safety as we identify new risks.Content Filtering
Prompts and completions are passed through a default configuration of Azure AI Content Safety classification models to detect and prevent the output of harmful content. Learn more about Azure AI Content Safety . Additional classification models and configuration options are available when you deploy an Azure OpenAI model in production; learn more .Model Specifications
Context Length131072
Quality Index0.80
LicenseCustom
Training DataOctober 2023
Last UpdatedFebruary 2025
Input TypeText,Image
Output TypeText
PublisherOpenAI
Languages27 Languages
Related Models
