News Summary

OpenAI’s recent launch of GPT-5 has received mixed reviews from software developers, who were told the model would serve as a ‘true coding collaborator’ excelling in generating high-quality code. Although GPT-5 is praised for its technical reasoning and cost-effectiveness, some developers find it less accurate compared to Anthropic’s competing models. The model’s verbosity settings affect its performance, sometimes leading to unnecessary code generation. Despite its cost advantages, GPT-5’s accuracy in reproducing scientific paper results was rated at 27%, compared to 51% for Anthropic’s premium model.