Workers AI 上线 Z.ai GLM-5.3 智能体编码模型,价格不变性能倍增
Workers AI - Z.ai GLM-5.3 now available on Workers AI
开发者注意:Workers AI 上线了 GLM-5.3,编码与智能体基准大幅提升,价格不变,性价比突出。若你已在 Workers AI 上使用 GLM-5.2,可考虑切换至 GLM-5.3 以获得更强能力,且无需调整预算。
@cf/zai-org/glm-5.3 is now available on Workers AI. It is Z.ai's flagship agentic coding model, built for long-running, tool-driven development workflows rather than single-turn chat.
@cf/zai-org/glm-5.3 现已在 Workers AI 上可用。它是 Z.ai 的旗舰级智能体编码模型,专为长时间运行、工具驱动的开发工作流而设计,而非单轮对话。
GLM-5.3 uses the same base model as GLM-5.2, with every gain coming from post-training. The results are substantial on coding and agentic benchmarks: Z.ai reports ↗ a 50% improvement over GLM-5.2 on its in-house Z.ai Code Bench, and calls GLM-5.3 the most capable open-weights model for coding. On public benchmarks, it scores 88.2 on Terminal Bench 2.1 (up from 81.0), 28.3 on Terminal Bench 3.0 — open-source state of the art, up from 4.6 — 66.9 on DeepSWE (up from 46.2), 78.1 on FrontierSWE (up from 67.5), and 42.5 on SWE-Marathon (up from 19.4). It is also the top-scoring model in Z.ai's comparisons on CyberGym for vulnerability discovery (84.5) and on long-horizon automation tasks like AutomationBench (48.2).
GLM-5.3 使用与 GLM-5.2 相同的基础模型,所有提升均来自后训练阶段。在编码和智能体基准测试上成果显著:Z.ai 报告 ↗ 在其内部 Z.ai Code Bench 上比 GLM-5.2 提升了 50%,并称 GLM-5.3 是编码能力最强的开放权重模型。在公开基准测试中,它在 Terminal Bench 2.1 上得分 88.2(从 81.0 提升),在 Terminal Bench 3.0 上得分 28.3——开源模型中的最优水平,从 4.6 大幅提升——DeepSWE 得分 66.9(从 46.2 提升),FrontierSWE 得分 78.1(从 67.5 提升),SWE-Marathon 得分 42.5(从 19.4 提升)。它还在 Z.ai 的 CyberGym 漏洞发现比较中得分最高(84.5),并在 AutomationBench 等长周期自动化任务中得分领先(48.2)。
The price-to-performance ratio is the compelling part. On Workers AI, GLM-5.3 costs the same as GLM-5.2 — $1.40 per M input tokens, $0.26 per M cached input tokens, and $4.40 per M output tokens — while roughly doubling GLM-5.2's scores on long-horizon benchmarks like SWE-Marathon, and improving them by more than 6x on Terminal Bench 3.0.
性价比是引人注目的亮点。在 Workers AI 上,GLM-5.3 的价格与 GLM-5.2 相同——每百万输入 token 1.40 美元,每百万缓存输入 token 0.26 美元,每百万输出 token 4.40 美元——而在 SWE-Marathon 等长周期基准测试上的得分大约是 GLM-5.2 的两倍,在 Terminal Bench 3.0 上提升了超过 6 倍。
GLM-5.3 requires the Workers Paid plan or prepaid AI Gateway credits.
GLM-5.3 需要 Workers 付费计划或预付费 AI Gateway 积分。
Use GLM-5.3 through the Workers AI binding (env.AI.run()), the REST API, the OpenAI-compatible endpoint, or AI Gateway.
通过 Workers AI 绑定(env.AI.run())、REST API、OpenAI 兼容端点或 AI Gateway 使用 GLM-5.3。
For more information, refer to the GLM-5.3 model page and pricing.
更多信息,请参阅 GLM-5.3 模型页面和定价。
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力