OpenAI’s GPT-5, released amidst high anticipation, showcases impressive capabilities in coding and logic but struggles with creativity. Touted as its “smartest and fastest model,” GPT-5 achieved notable benchmark scores, hitting 94.6% in math and 74.9% in real-world coding tasks. However, initial reviews were mixed, with users criticizing its creativity and suggesting it lacks the emotional depth of previous models. In particular, GPT-5’s storytelling prowess was rated low, producing stories that fell flat in creativity compared to competitors like Claude 4.1. While its coding abilities are deemed top-tier, offering clean and functional outputs, the model’s limitations in handling sensitive topics and its relatively small context window (400K tokens in total) make it less versatile in nuanced discussions. Overall, GPT-5 excels in technical tasks but seems less suited for users seeking a conversational or creative partner. OpenAI may need to iteratively enhance its performance to regain earlier users’ trust and satisfaction.
Source link
