OpenAI’s Deep Research smashes records for the world’s hardest AI exam, with ChatGPT o3-mini and DeepSeek left in its wake
Posted by:
Emma Walker
In a world where artificial intelligence is continuously evolving, recent developments have showcased remarkable progress in AI models’ accuracy. The top-ranking AI model, ChatGPT o3-mini, has seen a substantial improvement by 183% within a span of two weeks, now achieving up to 13% accuracy. Additionally, OpenAI’s Deep Research has set a new standard by achieving an impressive 26.6% accuracy on the world’s most challenging AI exam, Humanity’s Last Exam.
This groundbreaking achievement has stirred excitement within the AI community, as advancements in AI capabilities continue to push boundaries. With models like Deep Research showcasing proficiency in various subjects such as medicine, classics, and law, the potential for AI applications seems limitless. Although the road to surpassing the 50% accuracy mark on benchmarks like Humanity’s Last Exam may be daunting, the rapid progress indicates a promising future for AI technology.
As AI models like OpenAI’s Deep Research pave the way for enhanced analytical capabilities and efficient data processing, the horizon for technological advancements appears increasingly promising. Through continuous innovation and refinement, AI models are positioned to revolutionize industries and redefine the boundaries of what is achievable in the realm of artificial intelligence.
Recent Posts1
