Are you looking for smarter insights delivered straight to your inbox? Sign up for our weekly newsletters to receive the most relevant information for enterprise AI, data, and security leaders.
OpenAI’s Latest Updates
OpenAI has made GPT-4o, the large language model that powered ChatGPT prior to the recent launch of GPT-5, the default option for all paying users. This change applies to subscribers of ChatGPT Plus ($20 per month), Pro ($200 per month), Team ($30 per month), Enterprise, and Edu tiers. Users are no longer required to enable a “show legacy models” setting to access GPT-4o.
Furthermore, subscribers will now have access to a new “Show additional models” setting by default, which restores access to GPT-4.1, o3, and o4-mini, the latter two being reasoning-focused LLMs.
OpenAI’s CEO and co-founder, Sam Altman, announced this change on X, reassuring users that if GPT-4o is removed in the future, they will receive ample notice.
New Features in ChatGPT
Recent updates to ChatGPT allow users to select between “Auto,” “Fast,” and “Thinking” modes for GPT-5. While most users may prefer the Auto setting, the additional options provide more control for specific needs. The rate limits have been adjusted to 3,000 messages per week for GPT-5 Thinking, with additional capacity available for that mode.
The various models can be accessed through the “picker” menu at the top of the ChatGPT session screen, available on both web and mobile applications.
Challenges in AI Scaling
The landscape of enterprise AI is changing due to power limitations, rising token costs, and inference delays. To address these issues, we invite you to join our exclusive salon where leading teams are:
– Transforming energy into a strategic advantage
– Designing efficient inference for real throughput gains
– Unlocking competitive ROI with sustainable AI systems
Secure your spot to stay ahead: [https://bit.ly/4mwGngO](https://bit.ly/4mwGngO)
Reactions to GPT-5
The decision to revert to GPT-4o follows a tumultuous first week for GPT-5, which was launched on August 7 in four variants: regular, mini, nano, and pro, with optional “thinking” modes for more complex tasks. The rollout faced mixed reviews and infrastructure challenges, including a malfunctioning “autoswitcher” that misrouted prompts, inconsistent performance compared to GPT-4o, and user dissatisfaction over the abrupt removal of older models.
Altman’s latest update introduces new controls in the ChatGPT interface, allowing users to switch between “Auto,” “Fast,” and “Thinking” modes for GPT-5. The “Thinking” mode features a 196,000-token context window and is capped at 3,000 messages per week for paying subscribers. After reaching this limit, users can switch to the lighter “GPT-5 Thinking mini” mode. Altman indicated that these limits might change based on usage patterns.
However, GPT-4.5 remains exclusive to Pro users due to its high GPU costs. Additionally, Altman hinted at an upcoming personality adjustment for GPT-5, aiming for a “warmer” tone that is less polarizing than GPT-4o’s.
The company is also considering per-user customization as a long-term solution to address the emotional connections some users have developed with specific models. While these changes aim to ease frustrations over the transition to GPT-5 and the deprecation of older LLMs, they may also intensify the emotional attachments users have with these models.
Stay Informed with VB Daily
For those looking to impress their superiors, VB Daily offers valuable insights into how companies are leveraging generative AI. From regulatory changes to practical implementations, we provide the information you need to maximize ROI.
Subscribe Now
Read our Privacy Policy
Thank you for subscribing. Explore more VB newsletters here.