When will the AI price wars begin? (sancho.bearblog.dev)

AI pricing is at a pivotal moment, as major players like OpenAI and Anthropic continue to charge steep rates for their models, with costs reaching unsustainable heights for corporate users. While ChatGPT offers affordable subscription tiers for casual users, enterprise-level API rates can lead to exorbitant expenses, forcing companies to rethink their budgets. Despite Sam Altman's vision of “intelligence too cheap to meter,” current pricing trajectories suggest a different reality, with no signs of price reductions for older models as new releases come out. In a notable shift, the recent introduction of the open-weights model GLM-5.2 by the Chinese lab z.ai marks a significant turning point. Priced at $1 per million input tokens and $3 for output tokens—5 to 10 times cheaper than leading competitors—GLM-5.2 is poised to disrupt the market. Its open-access nature allows anyone with the necessary hardware to run the model independently, potentially promoting competition that could drive prices down further. As more providers adopt this model, the pressure will increase on OpenAI and Anthropic to reconsider their pricing strategies, potentially igniting an "AI price war" that could benefit consumers and transform the AI landscape.