Yuchen Jin Reports 1000x Drop in AI Token Prices

Yuchen Jin, research engineer at Databricks, posted a comparison of AI model costs. He stated that 1.5 years ago o1 Pro ran at $150 to $600 per million tokens. He added that GLM-5.3 Flash now runs at $0.15 to $0.50 per million tokens. Jin described the change as a roughly 1000x price collapse while claiming the newer model is more intelligent than o1 Pro. The post highlights intelligence per dollar as a measure of AI progress.