GPT-5 rollout stalls as legacy models return

OpenAI faces questions about whether GPT-5 is a real leap or a small upgrade.

OpenAI GPT-5 rollout hits turbulence as legacy models return

OpenAI launched GPT-5 with high expectations, but the rollout quickly ran into problems. The company rolled back some platform changes and restored access to GPT-4o for many users after an automatic router misrouted prompts and several performance hiccups were reported. Sam Altman acknowledged the hiccups in a Reddit AMA and on X, saying the rollout was 'a little more bumpy than we hoped for.'

User reports highlighted errors in math, logic, and coding tasks, while internal benchmarks did not always translate to real world use. OpenAI said it will adjust the model decision boundary, add a user option to trigger a thinking mode, and allow Plus users to keep using GPT-4o while it gathers more data. The incident has stirred questions about GPT-5's immediate value and has invited close scrutiny from developers, customers, and investors.

The episode shows how a major AI upgrade hinges not just on raw capability but on reliability and clear user experience. When a flagship model stumbles, trust thins and adoption slows, even for a platform that reaches hundreds of millions of weekly users.

OpenAI can still fix the routing, communicate better, and show real benefits with GPT-5. If the company can deliver faster, cheaper, and safer inference, it may recover momentum. If not, rivals that offer more stable, well understood tools will press their advantage and investors will demand sharper returns.

Highlights

The road to a confident GPT-5 remains long and uncertain.