DeepSeek Flash left a 1.8 Quadrillion Token Gap, AI model reveals "Impossible Triangle"

Insightful Beating AI News Update, OpenCode CEO Jay V stated that the DeepSeek Flash explosion has brought tremendous demand pressure. Over the past two weeks, the daily Token consumption of this model on OpenCode has increased from about 3 trillion to 18 trillion, a sixfold surge. Post-surge, OpenCode has been looking for suppliers who can handle traffic at a price close to the original, but it has been challenging to find service providers with both low prices and large-scale capacity.

Jay V believes that the challenge facing DeepSeek Flash is not insufficient demand but that low prices have unleashed demand far beyond expectations. It has simultaneously met the ability close to cutting-edge models and ultra-low costs, but once massive stable supply is introduced, the AI model market encounters the "impossible triangle": cheap, powerful, and with enough GPU capacity, three conditions that are difficult to satisfy simultaneously.

After the price increase of DeepSeek, OpenCode Go has also undergone multiple adjustments. The DeepSeek Flash quota was reduced from $60 to $15, then restored to $30, and an off-peak rate mechanism was introduced to align with the official pricing.