Arena.ai, which runs LMArena (formerly LMSYS Chatbot Arena), formally introduced 'Agent Mode' and the 'Agent Arena' leaderboard on June 4, 2026, measuring AI agents' real-world capabilities through actual user sessions. The system ranks frontier models based on how users actually put them to work executing complex tasks autonomously with tools, and an official demo video has also been released (YouTube).
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.