#预测市场 Seeing the data on Kalshi is still worth pondering. Gemini jumped from 30% at the beginning of the year to 86%, a significant reversal. In comparison, ChatGPT dropped from 41% to 8%, what does this shift indicate?

A trading volume of $14.08 million already shows that market participation is not low. The pricing of such prediction markets often reflects a relatively rational group consensus. As the settlement standard, the LM Arena ranking essentially compares the actual capabilities of the models. From the flow of funds, the market is re-evaluating the competitive landscape of major AI models, and Gemini’s policy adjustments and recent technological iterations seem to have gained market recognition.

However, long-term prediction markets also need to be cautious, as there is still time until the end of 2025, with many variables in between. Grok, although only accounting for 6%, as a latecomer, its subsequent performance is also within the scope of observation. The key still depends on actual model updates and market reactions; the numbers themselves are just the current pricing state.