OpenAI Shifts Gears: GPT-4o Returns, New Controls Emerge
OpenAI
GPT-4o
GPT-5
Large Language Models
AI
ChatGPT
Sam Altman
Generative AI
8
Iteration, Not Revolution
Media Hype
7/10
Real Impact
8/10
What is the Viqus Verdict?
We evaluate each news story based on its real impact versus its media hype to offer a clear and objective perspective.
AI Analysis:
While the shift is a response to immediate issues, the underlying AI technology is still developing rapidly, representing a moderate impact on the industry and significant continued hype due to OpenAI's market dominance.
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.