This week brought one of the biggest shake-ups we’ve seen on both the Arena and Artificial Analysis Index leaderboards.

The top two labs remain firmly in place, but Grok and Muse Spark made notable gains, putting pressure on Gemini and GLM. Even so, the performance gap to the top two is still significant.

What’s becoming increasingly important is the pace of intelligence improvement. It’s no longer just about reaching the top once. Labs need to iterate quickly to remain competitive.

More frequent model releases are also laying the foundation for continuous learning. If that trend continues, leaderboards may eventually update hourly as models evolve independently rather than through infrequent major releases.

We’ve already gone from quarterly releases to monthly ones. Could biweekly release cycles become the norm before the end of the year?