ViqusViqus
Navigate
Company
Blog
About Us
Contact
System Status
Enter Viqus Hub

OpenAI Shifts Gears: GPT-4o Returns, New Controls Emerge

OpenAI GPT-4o GPT-5 Large Language Models AI ChatGPT Sam Altman Generative AI
August 13, 2025
Viqus Verdict Logo Viqus Verdict Logo 8
Iteration, Not Revolution
Media Hype 7/10
Real Impact 8/10

Article Summary

OpenAI is attempting to course-correct following the initial rollout of GPT-5, which was met with significant user frustration due to infrastructure issues and inconsistent performance. The company has reversed its decision to make GPT-5 the default, reverting to GPT-4o for paying ChatGPT Plus subscribers. New settings have also been introduced, allowing users to select between ‘Auto,’ ‘Fast,’ and ‘Thinking’ modes for GPT-5, alongside a weekly message limit of 3,000 for the ‘Thinking’ mode. Furthermore, OpenAI plans to refine GPT-5’s personality, aiming for a more approachable tone while mitigating the polarizing effects of the previous iteration. This move represents a significant concession to user feedback, acknowledging the emotional investment many users have made in older models like GPT-4o and demonstrating a willingness to adapt its strategy. The company is grappling with the challenges of scaling AI infrastructure and managing user expectations, particularly as it moves towards increasingly complex and computationally intensive models.

Key Points

  • GPT-4o is now the default LLM for ChatGPT Plus subscribers.
  • New model selection controls ('Auto', 'Fast', 'Thinking') have been introduced for GPT-5.
  • Weekly message limits (3,000) are now in place for the ‘Thinking’ mode of GPT-5, with alternative lighter modes available.

Why It Matters

This news is critical for enterprise AI leaders because it signals OpenAI’s responsiveness to user concerns and highlights the ongoing challenges of deploying and scaling large language models. The initial rollout of GPT-5 demonstrated the potential for significant technical hurdles and user dissatisfaction, emphasizing the need for careful planning, robust infrastructure, and proactive engagement with the user community. Ultimately, this shift reflects a broader trend in the AI industry – a move towards more adaptable and user-centric development strategies.

You might also be interested in