AARENA匿名一对一对战，实时测试和比较各类 AI 模型。

4.8 (4)

审阅者 Daniel Nikulshyn·更新 2026年7月

Model Evaluation Benchmarking Community LLM Comparison Leaderboard Research

概览

AARENA 是一个让用户以匿名、实时方式让 AI 模型相互较量的平台。在评测过程中隐藏模型身份，鼓励用户仅凭输出质量做出公正判断，而非受品牌知名度影响。用户提交提示词后，会并排收到两个对战模型的回复，然后投票选出表现更优的一方。汇总后的结果有助于形成由社区驱动的排行榜，并揭示不同模型在各类任务中的表现差异。该工具非常适合希望对模型能力进行基准测试、探索替代方案，或想了解哪款 AI 最契合自身需求的研究人员、开发者和好奇用户。

主要功能

匿名模型对战
并排响应比较
用户投票系统
综合排行榜
支持多个AI模型
实时提示评估

价格

模型: Free
分类: 杰与拉常系安全目分
评分: 4.8 / 5 (4)

使用场景

盲测试验竞争性LLM

提交一个提示并比较两个匿名模型响应，并投票选出更好的输出，以评估没有品牌偏见的质量。

为研究基准测试模型

研究人员可以汇总多个提示的投票数据，以研究不同AI模型在不同任务上的表现，并生成社区驱动的排名。

为您的需求发现最佳模型

好奇的用户和开发者可以通过头对头测试模型，找出哪一个能最好地处理他们的用例，从而探索主流AI的替代方案。

在集成前验证模型选择

评估LLM用于产品的开发者可以通过AARENA运行真实提示，查看比较输出，并为购买或集成决策提供参考。

优点 & 缺点

优点

盲测试验减少品牌偏见
实时并排比较
社区驱动的排名
有助于基准测试多个模型
非技术用户也可访问

缺点

结果取决于主观投票
对模型内部机制的洞察有限
质量因提示类型而异

评测

4.8

4 个评分的平均值。

登录以留下评测。

Robert Ainsworth

May 3, 2026

Does the job

Pretty happy overall. Aggregated leaderboards just works and blind testing reduces brand bias. Limited insight into model internals can be annoying, but no dealbreakers — I'd recommend it to a friend without hesitating.

Omar Haddad

Mar 13, 2026

Use it every day

Honestly didn't expect to like it this much. Real-time prompt evaluation is exactly what I needed, and useful for benchmarking multiple models. but I reach for it almost every day now and it just clicks.

Diego Fernández

Aug 13, 2025

Skeptical, then convinced

I went in skeptical — most tools in this space overpromise. It actually delivers on side-by-side response comparison, and community-driven rankings caught me off guard. Quality varies by prompt type is why this isn't a perfect score, still, I'd recommend giving it a real trial.

Beatriz Costa

Jul 28, 2025

Years in this space

I've evaluated a lot of these over the years. What stands out here is side-by-side response comparison — handled better than most — and accessible to non-technical users. Worth the time if this is your use case.

问答

What are the main limitations of using AARENA for benchmarking?

Results rely on subjective user votes, which may vary by individual preference. The platform also does not expose internal model metrics, so users only see output quality, not technical performance details.

Asked by Diego Fernández · Sep 12, 2025

Can I compare more than two models at a time in AARENA?

AARENA supports side-by-side comparisons of two models per matchup, but users can submit multiple prompts to cycle through various model pairs and build a broader ranking over time.

Asked by Julia Steiner · Aug 31, 2025

How does AARENA prevent brand bias during model comparisons?

AARENA hides the identities of competing models in its interface, so users evaluate responses purely on quality, not on brand name. This blind setup ensures unbiased voting.

Asked by Olga Ivanova · Aug 20, 2025