Agents
Which AI gets real work done?
Agents take many steps on their own: browsing, running code, using apps. Here's who does it best, from real user sessions and from an open, repeatable test suite.
Agent Arena: top 20 models
Scored from 2,711,241 real agent sessions. Updated Nov 2, 2026.
1Claude Opus 5.5Max$6.82/task0.2002Gemini 4 CarbonHigh$1.73/task0.1663Claude Sonnet 5.5Max$3.94/task0.1464GPT-6 AstraMax$6.09/task0.1375Claude Fable 5.1Max$6.96/task0.1326GPT-6.1 SolMax$0.95/task0.1227Gemini 4 ArgonHigh$1.00/task0.1068GPT-6.1 TerraMax$0.79/task0.1029Muse Wonder FullMax0.09610Claude Fable 5High$3.92/task0.07811Claude Opus 5Max$5.62/task0.07612GPT-6 SolMax$1.74/task0.07013Kimi K3.1 ProMax$1.47/task0.04714GPT-5.6 SolxHigh$1.73/task0.04615Claude Sonnet 5High$1.11/task0.02816Grok 4.7xHigh$2.61/task0.02617Gemini 3.8 FlashHigh$0.76/task0.02618Muse WonderMax$1.43/task0.02519Muse Spark 1.3Max$0.45/task0.02420MiMo-V2.6-ProOPEN$0.16/task0.024
Source: Arena (formerly LMArena) leaderboard dataset, Agent Arena board, CC BY 4.0. The chip shows the median cost per task.
Agent setup × model
Exgentic Open Agent Leaderboard: the same 5 models run inside 5 agent setups on AppWorld, BrowseComp+, SWE-bench and Tau² (each counts a quarter). Share of tasks solved.
| Agent setup | DeepSeek V3.2 | Kimi K2.5 | GPT-5.2 | Claude Opus 4.5 | Gemini 3 Pro |
|---|---|---|---|---|---|
| Claude Code CLI | 42% | 30% | 39% | 67% | 56% |
| OpenAI Solo | 32% | 25% | 39% | 73% | 59% |
| SmolAgents Code | 41% | 42% | 38% | 66% | 56% |
| LiteLLM Tool Calling | 46% | 43% | 41% | 61% | 61% |
| LiteLLM Tool Calling with Shortlisting | 45% | 43% | 46% | 62% | 62% |
Best combo: OpenAI Solo with Claude Opus 4.5 (73%). These are older models: the dataset was last updated May 20, 2026. Source: Open Agent Leaderboard results (Exgentic), CDLA-Permissive-2.0.