According to DeepLearning.AI, independent lab Artificial Analysis has found that the cost of evaluating advanced chain-of-thought reasoning models, such as OpenAI o1, is rapidly escalating beyond the reach of resource-limited AI researchers. Benchmarking OpenAI o1 across seven widely used reasoning tests consumed 44 million tokens and incurred expenses of $2,767, highlighting a significant barrier for academic and smaller industry groups. This trend poses critical challenges for AI research equity and the development of robust, open AI benchmarking standards, as high costs may restrict participation to only well-funded organizations (source: DeepLearning.AI, June 18, 2025).