元記事の説明文
<p>Chinese AI lab <a href="https://z.ai/">Z.ai</a> released GLM-5.2 <a href="https://x.com/Zai_org/status/2065704919299235870">to their coding plan subscribers</a> on June 13th, and then yesterday (June 16th) released the full open weights under an MIT license. Similar in size to their previous GLM-5 and GLM-5.1 releases, this is 753B parameter, <a href="https://huggingface.co/zai-org/GLM-5.2">1.51TB</a> monster - with 40 active parameters (Mixture of Experts). GLM-5.2 is a text input only model - Z.ai have a separate vision family most recently represented by <a href="https://x.com/Zai_org/status/2039371126984360085">GLM-5V-Turbo</a>, but that one isn't open weights. GLM-5.2 has a 1 million token context window, up from GLM-5.1's 200,000.</p>
<p>The buzz around this model is strong.</p>
<p>Artificial Analysis, who run one of the most widely respected independent benchmarks: <a href="https://artificialanalysis.ai/articles/glm-5-2-is-the-new-leading-open-weights-model-on-the-artificial-analysis-intelligence-index">GLM-5.2 is the new leading open weights model on the Artificial Analysis Intelligence Index</a>.</p>
<blockquote>
<p><strong>GLM-5.2 is the leading open weights model on the Intelligence Index v4.1.</strong> At 51, it leads MiniMax-M3 (44), DeepSeek V4 Pro (max, 44) and Kimi K2.6 (43)</p>
</blockquote>
<p>They did however find it to be quite token-hungry:</p>
<blockquote>
<p><strong>GLM-5.2 uses more output tokens per task than other leading open weights models:</strong> the model uses 43k output tokens per Intelligence Index task, up from GLM-5.1 (26k) and above MiniMax-M3 (24k), Kimi K2.6 (35k) and DeepSeek V4 Pro (max, 37k)</p>
</blockquote>
<p>The model is also now ranked 2nd on the <a href="https://arena.ai/leaderboard/code/webdev">Code Arena WebDev leaderboard</a>, behind only Claude Fable 5. That leaderboard measures "front-end web development tasks, including agentic coding workflows". I'm impressed to see it rank so highly given the lack of image input, which I had incorrectly assumed was a key part of building a truly great frontend coding model.</p>
<p>I've been trying it out <a href="https://openrouter.ai/z-ai/glm-5.2">via OpenRouter</a>, which has it from 9 different providers, almost all of which are charging $1.40/million for input and $4.40/million for output. For comparison, GPT-5.5 is $5/$30 and Claude Opus 4.5-4.8 is $5/$25.</p>
<h4 id="excellent-pelican-disappointing-opossum">Excellent pelican, disappointing opossum</h4>
<p>GLM-5.1 gave me <a href="https://simonwillison.net/2026/Apr/7/glm-51/">one of my favorite pelicans</a> and my <a href="https://simonwillison.net/2026/Apr/7/glm-51/#opossum">all time favorite opossum</a> (for the prompt "Generate an SVG of a NORTH VIRGINIA OPOSSUM ON AN E-SCOOTER".) Interestingly, in both of those cases the model chose to return SVG wrapped in an HTML document that added additional animations using CSS.</p>
<p>Let's try GLM-5.2. For "Generate an SVG of a pelican riding a bicycle" I <a href="https://gist.github.com/simonw/5c989366b796f054d9ae1ad7e38dc03a">got this</a>:</p>
<p><img alt="It's a really good bicycle - all the right bits, spokes on the wheels, wheels and pedals rotating - and a very good pelican, red scarf, good beak, bobbing up and down. The feet don't stay on the pedals though." src="https://static.simonwillison.net/static/2026/glm-5.2-pelican.svg" /></p>
<p>It's a self-contained fully animated SVG, and the animations aren't broken! Often I'll see eyes falling off or wheels rotating independently of the bicycle but here everything works great. It's a very nice vector illustration of a pelican too. Very impressive.</p>
<p>Sadly, the NORTH VIRGINIA OPOSSUM ON AN E-SCOOTER did not come out <a href="https://gist.github.com/simonw/5913b56e3d0ba9a2ece75ce1471f87bb">nearly as well</a>:</p>
<p><img alt="Weird background gridlines, scooter is green and not very scooter like, possum is wearing a red safety helmet and has a hairy tail but is hardly recognizable as a possum. It's just bad." src="https://static.simonwillison.net/static/2026/glm-5.2-opossum.svg" /></p>
<p>This is such a step down from GLM-5.1! As a reminder, that possum looked like this:</p>
<p><img alt="This is so great. It's dark, the possum is clearly a possum, it's riding an escooter, lovely animation, tail bobbing up and down, caption says NORTH VIRGINIA OPOSSUM, CRUISING THE COMMONWEALTH SINCE DUSK - only glitch is that it occasionally blinks and the eyes fall off the face" src="https://static.simonwillison.net/static/2026/glm-possum-escooter.gif.gif" /></p>
<p>5.2 didn't even <em>try</em> to animate it.</p>
<p>Tags: <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/pelican-riding-a-bicycle">pelican-riding-a-bicycle</a>, <a href="https://simonwillison.net/tags/llm-release">llm-release</a>, <a href="https://simonwillison.net/tags/openrouter">openrouter</a>, <a href="https://simonwillison.net/tags/ai-in-china">ai-in-china</a>, <a href="https://simonwillison.net/tags/glm">glm</a></p>