ChatGPT Is Now The Fifth Top Global Website In Terms Of User Traffic: Report

But another recent report flagged concerns around OpenAI’s vulnerabilities.

ChatGPT is now the fifth top global website, according to a new report from RevealData.

OpenAI’s traffic exploded from 3.9 billion to 5.1 billion visits from February 2025 to April 2025 – a whopping 51.28 percent increase.

The numbers were calculated using data from SimilarWeb, RevealData said in a news release.

“ChatGPT has experienced exponential growth. While its uses are heavily varied, such as editorial assistance, data analysis, and image creation, it poses questions about how many people make AI a part of their lives. As AI becomes more powerful through new versions, it allows it to be used with less risk, and can add speed and accuracy to work, going from what was once niche, to something mainstream, and this can be seen with this huge explosion in web traffic,” said Wendell Jisa, founder and CEO of RevealData, in a statement.

Despite ChatGPT’s continuing dominance over global web traffic, there are concerns being raised about the vulnerability of the company’s technology.

Protect AI, which offers artificial intelligence and machine learning security, recently released a critical vulnerabilities report on OpenAI’s latest series of models, GPT-4.1, PT‑4.1 mini, and GPT‑4.1 nano.

According to OpenAI, these new models, released last month, “outperform GPT‑4o and GPT‑4o mini across the board, with major gains in coding and instruction following. They also have larger context windows—supporting up to 1 million tokens of context—and are able to better use that context with improved long-context comprehension.”

Protect AI used its red teaming product, Recon, to look for vulnerabilities among the new models.

Some of Protect AI’s findings included:

Testing resulted in 540 successful attacks across all three models, with 4.1 Mini having the most successful attacks at 193.
All models in this latest series were most susceptible to prompt injection attacks.
The models demonstrated “notable vulnerability to evasion attacks, which include prompts employing obfuscation techniques.”
GPT-4.1 Mini was tested as the most vulnerable overall, having a higher number of successful attacks.
The GPT-4.1 model attack testing showed its potential ability to assist in developing malware and performing harmful activities.

Protect AI summarized its findings as: “GPT-4.1 with its enhanced code writing capabilities is expected to find great adoption in building Agentic systems. However, our vulnerability analysis reveals that all three models in the series remain susceptible to generating harmful content across a range of harmful and illegal topics.”

MES Computing contacted OpenAI for comment.