GPT-4.5
FrontierMultimodalAPI Available
Benchmark Scores
88/100MMLU90.8
HumanEval88.5
Math81.2
Reasoning91.3
Coding88
About GPT-4.5
GPT-4.5 is OpenAI's largest and most knowledge-rich model to date. It excels at creative tasks, nuanced understanding, and tasks requiring broad world knowledge. It serves as a research preview of next-generation capabilities before GPT-5.
Strengths
- +Superior creative writing
- +Excellent emotional intelligence
- +Deep world knowledge
- +Strong agentic reasoning
Weaknesses
- -Very expensive
- -Slower than GPT-4o
- -Overkill for simple tasks
Pricing
Per 1M tokens
Input$75.00
Output$150.00
Quick Facts
- Context Window
- 128K
- Parameters
- Undisclosed
- Release Date
- 2025-02-27
- Category
- Frontier
Compare Models
See how GPT-4.5 stacks up against other models