OpenAI has launched its biggest and ‘best’ basic model to date, GPT-4.5. The Microsoft-backed AI startup claims that compared to previous models, GPT 4.5 improves on the ability to recognise patterns, make connections and generate creative insights without the need for reasoning skills.
OpenAI says it has used a scaling unsupervised learning paradigm for GPT-4.5, resulting in the model’s broader knowledge and deeper understanding of the world, while reducing hallucinations and increasing reliability.
In a post on X announcing the launch of GPT-4.5, Sam Altman said, “it is the first model that feels like talking to a thoughtful person to me. i have had several moments where i’ve sat back in my chair and been astonished at getting actually good advice from an AI.”
While GPT 4.5 isn’t a reasoning model like the o3 Mini or DeepSeek R1, Altman says there’s a “magic” to it that he hasn’t seen before.
GPT 4.5 doesn’t currently support multimodal features such as voice mode, video and screen sharing. However, OpenAI has indicated that these features may be coming in a future update.
Who can use GPT 4.5? What is its use case?
GPT 4.5 is currently only available to ChatGPT Pro users today. It will be rolled out to Plus and Team users from next week, with Enterprise and Edu users following the week after.
OpenAI says GPT 4.5 has a broader knowledge base and improved ability to track user intent, along with a higher EQ or emotional quotient.
All of this, the company claims, makes GPT 4.5 useful for tasks such as improving writing, programming and solving practical problems with reduced hallucination.
AI race heating up:
The artificial intelligence race is heating up at the moment, with companies launching one new offering after another since the sudden rise in popularity of China’s DeepSeek. This year alone, OpenAI has rolled out its o3 Mini reasoning model (announced in December), launched an AI agent to help with workflows called Deep Research, and now the introduction of GPT 4.5.
Earlier this month, Elon Musk’s xAI also rolled out a Grok 3 frontier model, a reasoning model and a Deep Search agent.Meanwhile, Google rolled out its Gemini 2.0 Flash reasoning model, while Amazon-backed Anthropic added the new 3.7 Sonnett model to Claude AI.
Meanwhile, Google rolled out its Gemini 2.0 Flash reasoning model (similar to o3 Mini or DeepSeek R1), while Amazon-backed Anthropic added the new 3.7 Sonnett model to Claude AI.
