Agent Arena | AI Agent Performance Leaderboard

Agent Arena🏆Overall

Dynamic ranking of models on how well they orchestrate tools for real-world agentic tasks, based on signals like tool reliability, task completion, and steerability.

Sep 30, 2026

2,022,521 sessions

48 models

Pareto Frontier

Net Improvement x Cost/task from real Agent Mode tasks

CtrlScroll

Pareto Optimal Models

Net Improvement x Cost/task

Claude Fable 5.1 (Max)

Anthropic · Proprietary

+14.55%

$4.42/task

Claude Opus 5.5 (High)

Anthropic · Proprietary

+13.78%

$1.56/task

GPT 6 Sol (Max)

OpenAI · Proprietary

+10.65%

$0.89/task

Gemini 4 Argon (High)

Google · Proprietary

+7.92%

$0.65/task

Hy4 preview

Tencent · Apache 2.0

+4.24%

$0.29/task

Deepseek V4.1 Flash (Max)

DeepSeek · MIT

+4.01%

$0.10/task

MiMo V2.6 Pro

Xiaomi · MIT

+3.17%

$0.10/task

GPT 6 Luna (Max)

OpenAI · Proprietary

+1.68%

$0.06/task

MiMo V2.6 Flash

Xiaomi · MIT

-0.57%

$0.04/task

Mimo V2.5 Pro

Xiaomi · MIT

-7.23%

$0.04/task

Cost per Task

Median cost per task (last 14 days)

1Claude Fable 5.1 (Max)$4.42/task+14.55%

2Claude Opus 5.5 (High)$1.56/task+13.78%

3GPT 6 Astra (Max)$2.80/task+12.18%

4GPT 6 Sol (Max)$0.89/task+10.65%

5Claude Opus 5 (High)$2.71/task+8.76%

6Claude Opus 5 (Max)$3.71/task+8.55%

7Claude Fable 5 (High)$2.13/task+8.37%

8Gemini 4 Argon (High)$0.65/task+7.92%

9GPT 5.6 Sol (xHigh)$1/task+7.05%

10Claude Opus 4.8 (High)$1.35/task+6.92%

11Claude Sonnet 5 (High)$0.94/task+4.81%

12Hy4 preview$0.29/task+4.24%

13Muse Spark 1.3 (Max)$0.42/task+4.19%

14GPT 5.5 (xHigh)$1.10/task+4.12%

15Deepseek V4.1 Flash (Max)$0.10/task+4.01%