AI models are becoming outdated at record speed — so what does that mean for the tokens that companies are spending millions on? (www.techradar.com)

AI models are evolving at an unprecedented pace, with new versions frequently replacing their predecessors, according to a report from Vercel’s AI Gateway. This rapid turnover is shifting token spending patterns, as most tokens are now being allocated to models that have been available for less than four months. The report highlights a significant movement towards open-weight models, which have captured 56% of token volume despite accounting for only 14% of the spending. This trend indicates a growing preference for tunable, open-source options over traditional closed models, exemplified by the rising success of TypeSafe AI’s Jev model. The implications of this rapid churn in AI models are considerable for the industry. As token prices continue to decline—down 23.2% over the last three months—companies may reevaluate their strategies for token procurement in light of the accelerating pace of innovation. The shift towards open-weight models not only offers cost-effective alternatives but also promotes greater flexibility for enterprises looking to adapt to the latest advancements. This ongoing transformation sparks a critical debate within the AI/ML community about whether the swift development cycle may hinder long-term stability and strategic planning in AI applications.