Mystery AI Model 'GPT-2' emerges, sparking OpenAI speculation

GPT-2 has unexpectedly appeared on a popular AI benchmarking website, LMSys
An undated image of ChatGPT AI. — Freepik
An undated image of ChatGPT AI. — Freepik

A new and powerful large language model (LLM) called ‘GPT-2’ has unexpectedly appeared on a popular AI benchmarking website, LMSys. The model's advanced capabilities and sudden arrival have ignited speculation that it could be a secret project from OpenAI, the renowned artificial intelligence research lab.

Details surrounding GPT-2 are scarce. It was added to the LMSys platform with minimal information, leaving researchers and enthusiasts scrambling to assess its abilities. Initial benchmarks suggest GPT-2 performs at an exceptional level, rivalling or even surpassing OpenAI's current flagship LLM, GPT-4. Users have reported GPT-2 excelling in reasoning tasks, code generation, and even artistic endeavours like creating ASCII art.

Read more: What is ChatGPT's memory feature and how does it work?

OpenAI has remained quiet about GPT-2. The company has a history of using secretive tactics in the past, including the surprise launch of GPT-3 in 2020. Some experts believe GPT-2 might be a public test bed for new technologies OpenAI plans to integrate into its next-generation LLM. Others suggest it could be a competitor's model or even a hoax.

The emergence of GPT-2 has undoubtedly shaken up the AI research landscape. If confirmed as an OpenAI project, it signifies a significant leap forward in LLM development. Regardless of its origin, GPT-2's capabilities highlight the rapid evolution of artificial intelligence and its potential to revolutionise various fields.