Benchmark
Arena Agent Leaderboard
A leaderboard focused on the capabilities of AI agents that evaluates model performance in multi-turn tool calling and real-world task scenarios, providing a reference for the cross-model ranking of agent capabilities.
Timeline (last 90 days)
Oct 28
On the Arena Agent Leaderboard, Gemini 4 Carbon ranks second, behind Opus.
Unverified50%
Oct 28
Arena Agent Leaderboard is a leaderboard focused on AI Agent capabilities, evaluating model performance in multi-turn tool calling, task planning and execution closed-loop scenarios.
Unverified50%