Developers Say GPT-5 Is a Mixed Bag

Some developers have said that they still have mostly positive experiences with GPT -5. Jenny Wang, an engineer, investor and personal styling agent Altar Creator, told Ward that the model seems better to complete complex coding tasks in a shot than other models. He compared it to OPANA O3 and 4O, which he used for code generation and straight fixes “like formatting, or if I want to create an last point similar to what I already have,” said Wang.

In his examination of GPT -1, Wang states that he asked the model to create a code for a press page for its company’s website, including certain design elements that match the rest of the site. GPT -5 completed the task at a tech, where in the past, Wang had to correct his requests during the process. Although there was a significant error, though: “It was hallusted the URLs,” said Wang.

Another developer, who spoke on the condition of anonymity because their employer did not allow them to speak to the press, said GPT -5 had the skills to solve deep technical problems.

The developer’s current hobby project is writing a programmatic network analysis tool, which is a one that will require code isolation for the purpose of security. The developer explained, “I originally presented my project and some of the way I was considering, and GPT -5 it took all things and returned a few recommendations with a real time limit,” the developer explained. “I’m fascinated.”

A handful of OpenAI’s enterprise partners and customers with cursor, Windsarf and concepts have proven to publicly coding and rational skills of GPT -5. (Many comments in OpenAI have included its own Blog post Announce the new model). Perception Divide This is “fast, thoroughly and complicated than the other models we tested to do 15 percent better”

However, within a few days of the release of GPT -5, some developers were weighing online with complaints. Many said that the coding capabilities of GPT -5 were thought to be a sophisticated, most powerful model from the world’s Begest AI company.

“Openai’s GPT -5 is very good, but it seems something that was published a year ago,” Kiran Classen, a developer created by AI assistant for email inboxes, says. “Its coding capabilities remind me of Sonnet 3.5,” he added, referring to an ethnographic model launched in June 2024.

Founder of Startup Organization Amir Salefefendius, Says in a social media post That he is using GPT -5 in the cursor and it is “quite underlying” and it has found “it is especially bad in coding.” He said that the release of GPT -1 was felt like a “Lama 4 moment” referring to Meta’s AI model, which disappointed some of the AI community.

X, developer McKay Rigle Wrote That GPT -5 is a “extraordinary daily chat model”, but when it comes to coding, “I will still use Clod Code + OPAS.”

Other developers describe the GPT-5 as “complete”-the time is helpful, but often disturbed in its long-boobs. Wang, who was overall satisfied with the front -end coding project prescribed at GPT -5, said he noticed that the model was “more unnecessary. It could clearly bring a cleaner or short solution.” (Kapoor noted that the verbosty of GPT -5 can be adjusted, so that users may ask it to be less chat or to argue less in exchange for better performance or cheap prices)))