ChatGPT-5 is frighteningly intelligent.

ChatGPT's "GPT-5 Thinking" feature is considered by experts to be incredibly effective. Photo: Shiftdelete .

GPT-5, the latest version of OpenAI's big language model, is being hailed by experts as a superior search tool, with the ability to automatically filter information using its "GPT-5 Thinking" feature.

Simon Willison, co-founder of Django – one of the world's most popular web frameworks – shared his incredible experiences using GPT-5. He likened the tool to a personal "research assistant" thanks to its ability to search and synthesize information quickly and accurately.

In a blog post titled "GPT-5 Thinking Has Surprisingly Good Search Capabilities," Willison argues that this new AI technology is a search assistant that goes far beyond a typical chatbot.

He believes that GPT-5 has completely changed the way he approaches information, enabling him to perform complex search tasks quickly and yielding better results than manual searches in the same amount of time.

To prove his point, Simon Willison provided several real-world examples. He recounted how he researched the history of "bouncing conveyors" at Heathrow Airport by simply assigning the task to GPT-5 instead of doing the research himself. The tool quickly compiled detailed and accurate information.

GPT-5 Thinking is capable of identifying a building based solely on a blurry image captured by Willison while he is moving. Photo: Simon Willison.

Similarly, GPT-5 also demonstrated superior capabilities when asked to identify a building from a blurry photograph or clarify whether Starbucks sells "cake pops" in the UK.

The most notable example is when Willison tasked GPT-5 with researching the history of the caves at Exeter harbor, a highly specialized mission. This AI tool performed a massive amount of work, searching for and synthesizing information from numerous sources.

The results not only provided answers but also included interesting historical details and even suggested pre-written emails to contact local researchers. Willison noted that this capability was very useful, not only for satisfying curiosity but also for important work requiring accuracy and speed.

Alongside praise for its incredible capabilities, the launch of the GPT-5 also faced considerable criticism from the user community. Immediately after its announcement, many expressed disappointment, claiming the new model was "less intelligent" or "less creative" than the GPT-40. Numerous users complained that the GPT-5 had a dry, unexpressive "personality" compared to previous models.

Dissatisfaction escalated when OpenAI removed or hid older models. Faced with strong protests, Sam Altman, CEO of OpenAI, acknowledged the problems and committed to making adjustments. The company subsequently restored access to some older models for paying users and refined the "personality" of GPT-5 to make it more user-friendly.