GPT-4.5 (openai.com)

Article summary

The discussion revolves around the release of GPT-4.5, with commenters sharing their thoughts on its capabilities and comparing it to other models like o3-mini and Claude 3.7. Some users have access to GPT-4.5 through the OpenAI platform, while others do not, despite having a Pro subscription. The model's performance is debated, with some citing its improved coding abilities and others questioning its value compared to cheaper alternatives.

Main themes

• GPT-4.5 release
• Model comparison
• Coding capabilities
• Pricing and accessibility
• AI evaluation metrics

What commenters say

• GPT-4.5's naming scheme is incoherent and lacks self-awareness, reflecting a larger issue with the model's development.
• The model's performance is not significantly better than its predecessors or competitors, making its release underwhelming.
• The cost of using GPT-4.5 is prohibitively expensive for personal projects, making cheaper alternatives like o3-mini more appealing.
• Human evaluation tests for AI models are subjective and potentially misleading, prioritizing likability over accuracy and capability.
• The value of GPT-4.5 lies in its improved coding abilities, making it a useful tool for developers and programmers.
• The comparison between GPT-4.5 and other models is unfair due to differences in pricing and accessibility, with some models offering better performance at a lower cost.
• The use of human evaluation tests to measure AI performance is necessary to determine its usefulness and truthfulness, despite potential limitations.
• GPT-4.5's release represents a step backwards in terms of nuance and intelligence, with its responses coming across as polarizing and less warm than its predecessors.