Z.ai刚刚放出了GLM-5.2，MIT开源、100万上下文、跑分81.0，跟Claude Opus 4.8差距很小，做开源模型的得看看。
Z.ai推出GLM-5.2开源模型，采用MIT开放权重。该模型支持1M上下文窗口，在Terminal-Bench 2.1基准上得分81.0，仅比Claude Opus 4.8低几分。Perplexity CEO此前指出，中国已拥有最强开源模型DeepSeek，且美国开发者正基于其构建应用。GLM-5.2的发布进一步表明开源AI竞赛已非理论。
🎯
🎯 Vadim @VadimStrizheus Perplexity CEO on China catching up in AI: “Whatever you did to not let them catch up didn’t even matter. They ended up catching up anyway. What’s more dangerous is they have the best open-source model. And all the American developers are building on that.” That was DeepSeek. Now Z.ai just dropped GLM-5.2: • MIT open weights • 1M context • 81.0 on Terminal-Bench 2.1 • within a few points of Claude Opus 4.8 The open-source AI race is not theory anymore. Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 1 🔄 1 ❤️ 15 👀 2223 📊 3 ⚡