Agent Arena Leaderboard: AI Agent Rankings & Benchmarks

Agent Arena | Capabilities🏆Overall

Dynamic ranking of models on how they perform in real Agent Arena sessions, based on signals like task success, user satisfaction, and steerability.

Oct 19, 2026

2,665,220 sessions

56 models

Pareto Frontier

Net Improvement x Cost/task from real Agent Mode tasks

CtrlScroll

Pareto Optimal Models

Net Improvement x Cost/task

Claude Opus 5.5 (Max)

Anthropic · Proprietary

+20.35%

$6.17/task

GPT-6 Astra (Max)

OpenAI · Proprietary

+15.15%

$4.12/task

GPT-6.1 Sol (Max)

OpenAI · Proprietary

+13.72%

$0.75/task

DeepSeek-V4.1-Flash-20261017 (Max)

DeepSeek · MIT

+5.13%

$0.19/task

MiMo-V2.6-Pro

Xiaomi · MIT

+3.74%

$0.12/task

GPT-6 Luna (Max)

OpenAI · Proprietary

+2.86%

$0.08/task

MiMo-V2.6-Flash

Xiaomi · MIT

+2.54%

$0.06/task

GPT-6 Luna (Medium)

OpenAI · Proprietary

-1.04%

$0.02/task

GPT-6 Luna (Low)

OpenAI · Proprietary

-4.80%

$0.01/task

Cost per Task

Median cost per task (last 14 days)

1Claude Opus 5.5 (Max)$6.17/task+20.35%

2GPT-6 Astra (Max)$4.12/task+15.15%

3Claude Fable 5.1 (Max)$7.69/task+14.33%

4GPT-6.1 Sol (Max)$0.75/task+13.72%

5Claude Sonnet 5.5 (Max)$4.05/task+12.75%

6Gemini 4 Argon (High)$1.10/task+12.12%

7GPT-6 Sol (Max)$1.54/task+9.33%

8Claude Fable 5 (High)$3.28/task+8.37%

9Claude Opus 5 (Max)$5.35/task+8.08%

10GPT-5.6 Sol (xHigh)$1.60/task+6.59%

11DeepSeek-V4.1-Flash-20261017 (Max)$0.19/task+5.13%

12Claude Sonnet 5 (High)$1.22/task+4.06%

13Muse Wonder (Max)$1.58/task+3.77%

14MiMo-V2.6-Pro$0.12/task+3.74%

15Gemini 3.8 Flash (High)$0.79/task+3.69%