Summary: LMSYS Chatbot Arena adds Gemini 3 and GPT-5.2

Why It Matters: The platform has updated its leaderboard with several unreleased or next-generation AI models. This gives a first look at how new models from Google, OpenAI, and Anthropic rank against each other in human preference tests.

What Changed:

• Added several new "Gemini 3" model variants (Pro, Flash, and Thinking).
• Added "GPT-5.1" and "GPT-5.2" models to the rankings.
• Added "Claude 4.5" and "Claude 4" thinking models.
• Added "Grok 4.1" and "Grok 4" reasoning models.
• Updated the leaderboard rankings for over 290 different AI models.
• Removed the security "verify you are human" landing page to show the full data table.