Summary: LMArena launches Gemini 3 and Grok 4.1 rankings

Why It Matters: The leaderboard now includes unreleased or "frontier" AI models from Google, xAI, and Anthropic. This gives developers and founders a preview of the next generation of AI performance across text, coding, and vision tasks.

What Changed:

• Added rankings for new models including Gemini 3 Pro, Gemini 3 Flash, and Grok 4.1.
• Introduced a "Thinking" category for models like Claude 4.5 and Grok 4.1.
• Expanded the leaderboard to include specific tabs for WebDev, Search, and Video generation.
• Updated the "Arena Overview" with detailed stats for 293 different models.
• Added a "Follow us" link to their X (Twitter) account for frontier AI news.