#gpt-5

#gpt-5

When OpenAI launched its new GPT-5 model in August, the company bragged loud and hard about how GPT-5 is its "smartest, fastest, most useful model yet" and how interacting with it was like "chatting with a helpful friend with PhD‑level intelligence." When it comes to creative tasks like writing, GPT-5 immediately felt like a major step backward. But as I've tested the model more extensively, I've seen that it does excel at many pragmatic tasks like writing code and analyzing data.

Based on over 4,400 Java tasks, the report finds that depending on which of the four levels of reasoning capabilities that OpenAI now makes available, the overall quality of the code, especially in terms of the vulnerabilities generated, significantly improves. However, the overall volume of code being generated per task also substantially increases, which creates additional maintenance challenges for application developers that are not going to be familiar with how code might have been constructed in the first place.

OpenAI's launch of its long-awaited GPT-5 AI model turned out to be a bit of a dud. Those expecting a revolutionary change, something CEO Sam Altman promised outright months ago, were left sorely disappointed. In many ways, GPT-5 felt more like an iterative improvement, while a colder and less personable tone took aback those looking to foster an emotional relationship with the bot.

"We totally screwed up," Altman admitted in an interview with The Verge.

"The Trump administration is creating and sharing memes to make fun of deportations, showcasing how humor is used in political discourse."

Some developers report positive experiences with GPT-5, noting its improved ability to tackle complex coding tasks efficiently compared to previous models like o3 and 4o.

This week on The Vergecast, we're discussing the fallout of the messy GPT-5 launch and why the new model isn't living up to the hype.

The premium plan called ChatGPT Plus guarantees access during blackout periods and includes access to OpenAI's most advanced models, making it attractive for superusers.

The simple truth is that medicine is always multimodal. A physician's mind doesn't travel in a straight line, drifting from the patient's story to the CT image, lab values, and clues in a physical exam.

Altman expressed optimism for Generation Z, stating, 'I would feel like the luckiest kid in all of history,' despite acknowledging potential job displacement from AI.

OpenAI's GPT-5 signifies a major enhancement in AI intelligence, offering user-friendly interactions by routing tasks automatically to the most suitable model, making it accessible to all.

"We use Echo Chamber to seed and reinforce a subtly poisonous conversational context, then guide the model with low-salience storytelling that avoids explicit intent signaling," security researcher Martí Jordà said.

In mid-2023, querying GPT-5 for a recipe could consume several to 20 times more energy than previous models, indicating increased resource demands for advanced AI responses.