Scientists were able to train a competitor to o1 from OpeanAI for just $50

Researchers from Stanford and Washington Universities managed to train an artificial intelligence model capable of o1-level reasoning from OpeanAI for only $50. This was reported by TechCrunch.

The model is called s1, and it performs as well in tests as state-of-the-art reasoning models such as OpenAI's o1 and DeepSeek's R1. The s1 model is available on GitHub, along with the data and code used to train it.

The s1 development team reported that they started with a ready-made base model and then improved it through distillation, a process that involves extracting reasoning abilities from another AI model by learning from its responses. According to the researchers, 16 NVIDIA H100 GPUs were used to train s1 and took about 30 minutes.

The s1 model is based on a small, off-the-shelf AI model from Qwen, a Chinese AI lab owned by Alibaba, and trained on the Gemini 2.0 Flash Thinking Experimental reasoning model.

The researchers report that reasoning models can be derived from a relatively small dataset through a process called supervised fine-tuning (SFT), in which the AI model is explicitly instructed to mimic certain behaviors in the dataset.

The idea that a few researchers without millions of dollars in funding can still innovate in the field of AI is indeed fascinating. But s1 raises important questions about the commercialization of AI models, and the fact that anyone can now accurately reproduce a multi-million dollar model for little money.

In 2025, Meta, Google, and Microsoft plan to invest hundreds of billions of dollars in AI infrastructure, which will be partially used to train next-generation AI models. Distillation has proven to be a good method for cheaply replicating artificial intelligence capabilities, but it does not create new AI models that are better than those that already exist.