Sam Altman's Openai has released GPT-4.5, an upgraded version of the artificial intelligence model that runs ChatGpt, and is trying to stop the challenges from rivals like Elon Musk's Xai by selecting users on Thursday.
Described as Openai's “although the biggest and best model for chat”, GPT-4.5 was initially available as a “research preview” for a limited number of software developers and ChatGPT Pro subscribers.
GPT-4.5 does not have the tendency to be so-called “hastised.” This is a phenomenon in which AI models produce false responses to user prompts. According to Openai, hallucination rates for the upgraded model were only 37.1%, down from the last version of 61.8%.
“Early tests show that interaction with GPT ‑ 4.5 feels more natural,” Openai said in a blog post. “Its broad knowledge base, improved ability to follow user intent, and a larger “EQ” can help with tasks such as improving practical problems, programming, and solving them. ”
Altman described the GPT-4.5 as “the first model that feels like talking to thoughtful people,” but also tried to ease expectations about the launch.
He admitted there was an update “Not crushing benchmarks” and “a huge and expensive model.”
Backed by Microsoft, Openai faces increasingly tough competition from many rivals, from Google's Gemini Ai to Xai's Grok and China-based Deepseek.
CHATGPT is the world's leading chatbot service with over 400 million weekly active users.
Earlier this month, Musk unveiled the Grok 3 models. This scored higher in mathematics, science and coding tests than Openai's GPT-4o, Google's Gemini, Anthropic's Claude and Deepseek's V3 models.
The battle between Musk and Altman also expanded to federal courts. Musk filed an anti-trust lawsuit against Openai, seeking an injunction to block Altman's plan to convert startups into commercial companies.
Musk co-founded Openai with Altman, but left the startup after disagreement about its long-term plans. The two billionaires then become bitter rivals and frequently exchange national humiliations.
Elsewhere, Mark Zuckerberg's Meta announced this week that it will launch its own standalone app for its AI chatbot, which competes directly with ChatGpt.
In response to the announcement, Altman joked that Openai could challenge his Facebook and Instagram parents to his lawn.
“It might be okay, but I'll do a social app.” Altman wrote X. “If LOL Facebook is about to come to us and we just reverse them, it would be very interesting.”
