Google Introduces AI Reasoning Model Rivaling OpenAI

Google is developing an AI reasoning model capable of solving complex math and programming problems, using the chain-of-thought technique. 

This latest advancement by Google aims to directly rival OpenAI’s o1 reasoning model. According to people familiar with the matter, multiple teams within Alphabet’s Google have been working secretly on the development of innovative AI systems with complex problem-solving capabilities, such as math and coding.  

Chain-of Thought for Enhanced Problem-Solving 

One of the key components of this new AI reasoning model is the use of technique dubbed as “chain-of-thought prompting.” This method allows the AI model to think about the answer it will be generating to take into consideration some ideas related to the topic and then give a more accurate and elaborated answer. 

By using this approach, the AI model will be able to mimic human reasoning. It is worth noting that Google is not the first company to adopt this method, as OpenAI has already used it for its AI models, including the latest version, the o1 model. 

When reached for a comment, Google declined to give further details about the development, but the same insiders confirm that work is underway. AI researchers from both Google and OpenAI are competing in efforts to take reasoning AI to the next level. 

A Never-Ending Rivalry 

Since the Microsoft-backed company introduced its AI reasoning model, known as “Strawberry” in September, researchers of DeepMind have been put under pressure to push the AI advancements further. However, the concerns of Google falling behind following the release of ChatGPT has diminished over the past few months, as the search-engine company has revealed several AI developments. 

The competition between the two giants is constantly growing, as many are wondering about the future of the technology field, as OpenAI is continuously growing and gaining popularity to the extent that some industry observers believe that AI could challenge the dominance of Google. 

Although Google has been very cautious in its developments compared to its rival, it remains a key player in the AI landscape. 

Oren Etzioni, an AI researcher and founder of TrueMedia.org, a nonprofit working to combat political disinformation, weighed in on the competition. 

“Technically it’s always been the case that Google’s capabilities were top-notch. They were just more conservative in rolling things out… It’s a marathon, and it’s anybody’s race to win,” Etzioni said. 

Continuous Developments 

At the more moderate pace of development, Google reports significant progress. In July, for instance, the company released AlphaProof, an AI in mathematics that reasons like a wizard, and AlphaGeometry 2, an updated model better adapted to solving geometric problems. Both systems performed well in the International Mathematical Olympiad, according to Google, they solved four out of six problems posed. 

During its I/O conference in May, Google also showcased Astra-an AI assistant capable of utilizing the camera of a phone to identify objects surrounding its user. The company hinted that some of Astra’s features could be integrated into its flagship AI model, Gemini, later this year. 

“Advanced mathematical reasoning is an important capability for modern AI,” said Google DeepMind CEO Demis Hassabis in July, further underlining the commitment of Google to stay ahead in the evolving landscape of AI. 

Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Tech sections to stay informed and up-to-date with our daily articles.