https://vibecoding.rest/tags/z.ai/Recent content in Z.ai on Vibe CodingHugoenhttps://vibecoding.rest/models/glm-5-2/Mon, 01 Jan 0001 00:00:00 +0000https://vibecoding.rest/models/glm-5-2/<p>GLM-5.2 is Z.ai&rsquo;s (formerly Zhipu AI) flagship open-weight model — a 753-billion-parameter mixture-of-experts design released under the MIT license, with a 1M-token context window and full downloadable weights.</p> <h2 id="why-it-stands-out">Why it stands out</h2> <ul> <li><strong>MIT-licensed, 753B-parameter MoE weights</strong>, fully self-hostable with no usage restrictions.</li> <li><strong>1M-token context window</strong> with up to 131K tokens of output in a single response.</li> <li><strong>Beats GPT-5.5 on FrontierSWE</strong> at roughly a sixth of the cost, per Zhipu&rsquo;s own benchmarking.</li> <li><strong>Selectable High/Max reasoning modes</strong>, letting you trade latency for depth on harder multi-step coding tasks.</li> </ul> <h2 id="good-for">Good for</h2> <p>Teams that want a genuinely open, frontier-tier coding model to self-host, fine-tune, or run at scale without per-token lock-in.</p>