OpenAI Releases Major Update With GPT-5
OpenAI released GPT-5 last week with some notable improvements and our team watched the livestream. Here's what we think is most important to know about the update.
What we are excited about:
- Free access to their most advanced model with reasoning (though it'll likely be heavily limited as 4o access is now)
- New voice mode available to all users
- Single model experience - no more confusion about which version to choose
- Fewer hallucinations although we’ll believe it when we see it
- Improved handling of difficult or controversial topics
- Much larger memory in chats - about 3 times the current context window size
What caused our team some concern:
- Increased focus on health advice, which raises questions about dependence on AI in high-risk situations
- Enhanced user memory including Gmail and calendar connections, which presents serious privacy concerns for educators
- In the showcase of Study Mode for learning Korean the bot just gave students the answer instead of guiding them to discover it on their own
We haven't had time to test these claims yet, so we'll be trying it out over the coming weeks in our workshops. So we will check back-in with our honest assessment once we've had some hands-on experience.
You can learn more about OpenAI's GPT-5 release here.
