OpenAI’s GPT-5 Launch Sparks Mixed Reactions Amid Benchmark Dominance

OpenAI unveiled GPT-5 last week, positioning it as its "smartest, fastest, most useful model yet." The AI excels in technical tasks, boasting benchmark scores of 94.6% on math tests and 74.9% on real-world coding challenges. Sam Altman likened its capabilities to having a team of PhD-level experts on demand.

Despite its prowess in logic-driven tasks, GPT-5 faces criticism for its limited creative output and small context window. Early adopters flooded Reddit with complaints, labeling the model "horrible" and "underwhelming." The backlash prompted OpenAI to reintroduce GPT-4o as a legacy option.

The divide highlights a tension between raw performance and user experience. While GPT-5 dominates benchmarks, its inability to win hearts underscores the challenges of balancing technical excellence with human-centric design.