OpenAI recently unveiled the next-generation large language model, GPT-5, which now powers ChatGPT and other AI tools from the AI maker. This is huge news for educators and their students because ChatGPT is the most used AI tool and now is claiming to have the most powerful public-facing AI model.
The launch of GPT-5 has met with some criticism, however, with the consensus seeming to be that it is more of a routine upgrade type of improvement than the seminal moment in AI development that was promised—and that’s been my personal experience as well.
Here’s a closer look at everything teachers need to know about GPT-5.
What Is GPT-5?
The first version of ChatGPT, which launched on November 30, 2022, was powered by GPT-3.5. The capabilities of GPT-3.5 may seem quaint by today’s standards but were impressive enough to launch the age of AI in education. OpenAI followed that release with GPT-4 in March 2023, and later rolled out models such as GPT-4o in May 2024 and GPT-4.5 in February of this year.
OpenAI suggested that GPT-5’s long-delayed release would mark a new era in AI capabilities. “It’s a significant step along the path of AGI,” said OpenAI CEO Sam Altman in a press briefing about GPT-5. He added that it was like having access to a PhD-level expert on any topic in your pocket.
Many find these statements to be hyperbolic, however.
How Are People Reacting To GPT-5?
Many people seem disappointed with the tool’s capabilities at launch. A story in Axios declared that GPT-5 “landed with a thud,” while an article in The Conversation wondered if AI has plateaued. Some users were upset that they no longer had access to previous models, and that GPT-5 seemed to lack the personality of some previous models.
In my use, I haven’t noticed lag time or a shift in tone, but I haven’t seen any major improvements, either. In fact, I’m a little embarrassed to admit that if no one had told me the model had been updated, there’s almost zero chance I would have noticed.
That said, the tool does have some specific improvements. It’s supposed to hallucinate less, be less of a sycophant, and is designed to choose the best type of model for your query, so more complex questions will be routed to a reasoning model that might take longer to answer, but should provide a better answer. Finally, GPT-5 is designed to be better at coding.
The disappointment with GPT-5 was not universal. Tom’s Guide found that GPT-5 easily outperformed Google’s top model, Gemini 2.5, and offered a clear improvement when GPT-5 went head-to-head with GPT-4.
How Does GPT-5 Handle Class Writing Assignments?
For an idea of what GPT-5 looks like in action, here’s a prompt I shared with it.
My prompt: Write a one- to two-paragraph response paper about a piece of writing that inspired you. This can be a newspaper or magazine article, a poem or work of fiction, or any writing you have been drawn to.
ChatGPT said:
One piece of writing that has deeply inspired me is James Baldwin’s essay “Notes of a Native Son.” Baldwin’s ability to weave personal narrative with broader social commentary creates a work that is both intimate and universal. He writes with unflinching honesty about his father’s death, the racial tensions of 1940s America, and his own experiences navigating a hostile world. What struck me most is Baldwin’s mastery of balance—his prose is lyrical yet precise, emotional yet clear-eyed, and his anger is tempered by an underlying search for understanding. Reading this essay feels like being invited into a deeply personal conversation that still resonates with urgent relevance today.
What inspired me most was Baldwin’s refusal to simplify the truth. He does not shy away from contradictions—he acknowledges love for his father alongside bitterness, pride in his identity alongside frustration with injustice. This complexity encouraged me to think about my own writing differently, to embrace nuance rather than force neat resolutions. Baldwin’s essay showed me that the most powerful writing often leaves the reader unsettled, still turning over questions long after the final sentence. It’s a reminder that great writing doesn’t just reflect the world—it challenges us to see it, and ourselves, more clearly.
To those who regularly encounter AI writing, GPT-5 has a familiar writing style. This passage is technically proficient writing to be sure, but suffers from that same old AI lifelessness with phrases such as “his prose is lyrical yet precise, emotional yet clear-eyed.” These phrases sound nice but convey little meaning and are more generic than most human writers could muster in a lifetime of writing.
What Does This All Mean For Teachers?
For those of us who don’t teach code, GPT-5’s capabilities probably won’t change much, but it’s still something to keep an eye on. It might be marginally better at helping a teacher who uses it for lesson planning, or a student exercise involving AI might be better put together.
On the other hand, students who use AI to cheat might get slightly better papers that, in theory, could be more difficult for educators to spot. Ultimately, I don’t think it’s going to make a big difference.
Of course new uses cases are likely to emerge, but right now I don’t see GPT-5 disrupting the education space in any kind of major way, positive or negative. For me personally, this is both a relief–and also a little disappointing.