Skip to content
Trending storyDeveloping

Arena ranks Claude Haiku 5.5 (High) at #27 in Agent Arena

1 article1 sourcesince Oct 14Last article 1h ago ·

Overview

AISummary of 1 article

Arena ranks Claude Haiku 5.5 (High) from Anthropic at #27 in Agent Arena, with a +1.48% net improvement across more than 20.5K real-world agentic sessions.

Its best category is Code, where it ranks #15 with +4.83% net improvement, according to Arena.

Arena reports a median cost of $0.09 per task for Haiku 5.5, placing it just outside the Agent Arena Pareto frontier. Arena says this is 86% less per task than GPT-5.4 (High), which costs $0.63.

Written by AI from the articles below · updated Oct 14, 12:34 PM ET

Check the sources:

Latest developmentOct 14, 12:27 PM ET
Claude Haiku 5.5 ranks #27 in Agent Arena at $0.09 per task

Article timeline

The articles in this story. Times are ET.

Oct 14
  1. Arena.aiOfficial
    Claude Haiku 5.5 ranks #27 in Agent Arena at $0.09 per task

    AIClaude Haiku 5.5 (High) from Anthropic ranked #27 in Agent Arena across more than 20.5K real-world agentic sessions, with a +1.48% net improvement. Its strongest category is Code, where it ranks #15 with +4.83%, and its median cost is $0.09 per task, just outside the Pareto frontier. Anthropic's Haiku 5.5 costs 86% less per task than GPT-5.4 (High) at $0.63, according to Arena.

    Image from @arena's post

Heat trend

Not enough continuous observations to show a trend yet.