What micro1 changed

• Removed a link to an external "End of Q3 update" post on X (formerly Twitter) from the main page.

The removal of a quarterly update link suggests the company is moving past its previous reporting cycle or cleaning up older promotional content to focus on current initiatives.

• Updated the "Realm: Legal reasoning benchmark" rankings.
• Added "Gemini 4 Argon (high)" to the second-place spot with a score of 66.5%.
• Removed "Claude Fable 5.1 (max)" from the top three list.
• Moved "Claude Sonnet 5.5 (max)" from second place to third place.

The company updated its legal reasoning leaderboard to include a new high-performing model. This shows they are actively testing the latest AI models to stay relevant as a source for AI evaluation and data.

• Updated the PersonalAgentBench leaderboard with new top performers: Dots (80%) and Hark (70%).
• Increased Gemini Spark's score from 60% to 65% while moving it to third place.
• Removed Instinct and Grok Bot from the homepage leaderboard summary.
• Added a link to an "End of Q3 update" from the founder on X (formerly Twitter).

The company updated its benchmark results to show new AI agents (Dots and Hark) outperforming previous leaders like Gemini Spark. This positions their evaluation platform as the source for tracking the newest, high-performing agents.