<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: song chou</title>
    <description>The latest articles on DEV Community by song chou (@amber_li_b5d56b3d448b7c85).</description>
    <link>https://dev.to/amber_li_b5d56b3d448b7c85</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4256995%2Fc88cc068-0e9a-4930-a6f7-3325f4016060.png</url>
      <title>DEV Community: song chou</title>
      <link>https://dev.to/amber_li_b5d56b3d448b7c85</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/amber_li_b5d56b3d448b7c85"/>
    <language>en</language>
    <item>
      <title>Three Common AI Model Review Sites</title>
      <dc:creator>song chou</dc:creator>
      <pubDate>Sat, 21 Nov 2026 09:30:03 +0000</pubDate>
      <link>https://dev.to/amber_li_b5d56b3d448b7c85/three-common-ai-model-review-sites-2mn5</link>
      <guid>https://dev.to/amber_li_b5d56b3d448b7c85/three-common-ai-model-review-sites-2mn5</guid>
      <description>&lt;p&gt;One site tells you the numbers. One tells you which model wins a blind vote. The third — which I built — tells you whether people think it's worth the money.&lt;/p&gt;

&lt;p&gt;If you've ever picked an AI model because the benchmarks said so, and then regretted it a week later, you know those three things are not the same. Here's the difference.&lt;/p&gt;




&lt;h2&gt;
  
  
  1. Artificial Analysis — the numbers
&lt;/h2&gt;

&lt;p&gt;Artificial Analysis is the closest thing the AI world has to a spec sheet.&lt;/p&gt;

&lt;p&gt;It measures intelligence, speed, and price across hundreds of models, updates constantly, and doesn't have an opinion. That's the point, and it's also the limit.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's good for:&lt;/strong&gt; the honest answer to "which model is objectively the strongest right now."&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's not:&lt;/strong&gt; a recommendation. A model can top the intelligence index and still be a miserable thing to use for your actual job. Artificial Analysis won't tell you that, because it's not trying to.&lt;/p&gt;

&lt;p&gt;Website: &lt;a href="https://artificialanalysis.ai" rel="noopener noreferrer"&gt;https://artificialanalysis.ai&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fk1cq0q6eg9ou8op3pm74.png" alt=" " width="800" height="359"&gt;
&lt;/h2&gt;

&lt;h2&gt;
  
  
  2. LMArena — the blind vote
&lt;/h2&gt;

&lt;p&gt;LMArena shows you two anonymous answers and asks which is better. Millions of votes later, you get an Elo ranking that reflects what humans actually prefer.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's good for:&lt;/strong&gt; the answer to "which one feels better to talk to."&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's not:&lt;/strong&gt; the answer to "which one should I pay for." A blind vote can't tell you if the model that wins is also the one that burns your monthly quota in an afternoon. Preference is not the same as value.&lt;/p&gt;

&lt;p&gt;Website: &lt;a href="https://lmarena.ai" rel="noopener noreferrer"&gt;https://lmarena.ai&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3bngs36w6dyfsvm5crtr.png" alt=" " width="799" height="458"&gt;
&lt;/h2&gt;

&lt;h2&gt;
  
  
  3. highworth.ai — the "worth it" score (I built this one)
&lt;/h2&gt;

&lt;p&gt;Highworth flips the question from "which is better" to "was it worth the money."&lt;/p&gt;

&lt;p&gt;Every model gets a score from real people who've used it — pulled from Hacker News, Reddit, V2EX, and Bilibili, with every quote linked back to its source. The result is a percentage: what share of people thought it was worth paying for.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's good for:&lt;/strong&gt; the honest answer to "should I spend on this."&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it's not:&lt;/strong&gt; a benchmark. The sample is smaller than the big arenas, and the audience skews technical. A score from three reviews is not the same as one from fifty, and we say that on the page.&lt;/p&gt;

&lt;p&gt;Website: &lt;a href="https://highworth.ai" rel="noopener noreferrer"&gt;https://highworth.ai&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  &lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fh1qwajhg8wuafd6tb43u.png" alt=" " width="800" height="526"&gt;
&lt;/h2&gt;

&lt;h2&gt;
  
  
  So which one do you use?
&lt;/h2&gt;

&lt;p&gt;The boring but true answer: all three, in order.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Start with Artificial Analysis to see what's objectively strong.&lt;/li&gt;
&lt;li&gt;Check LMArena to see what people actually like using.&lt;/li&gt;
&lt;li&gt;Then check highworth to see if people thought it was worth paying for.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;One tells you what's best. One tells you what's preferred. One tells you what's worth it. You usually need all three before you open your wallet.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>showdev</category>
      <category>discuss</category>
      <category>productivity</category>
    </item>
  </channel>
</rss>
