OpenAI CEO Sam Altman reassured users of the new GPT-5 model, which was released on Thursday, that it will improve over time. During a Reddit “ask me anything” session, multiple users complained that GPT-5 wasn’t working as well as its predecessor, GPT-4.
Altman attributed the initial issues to the new router system, which was not functioning properly when it was first rolled out. He promised that GPT-5 will seem “smarter” starting today, with improved autoswitching and more transparent decision-making processes.
To address user concerns, Altman announced plans to double rate limits for Plus subscribers as the rollout continues. This will give users a chance to test the new model without worrying about running out of monthly prompts.
The CEO also acknowledged the “chart crime” incident during Thursday’s presentation, where a misleading chart was presented, leading to numerous jokes and memes. Altman promised to work on fixing these issues and ensuring that users get accurate results.
Despite initial concerns, many reviewers praised GPT-5, including Simon Willison, who said it performed well in early tests. However, Altman emphasized the importance of continued feedback and improvement, promising to keep listening to user input as the model evolves.
Source: https://techcrunch.com/2025/08/08/sam-altman-addresses-bumpy-gpt-5-rollout-bringing-4o-back-and-the-chart-crime