大模型
www.aibase.com ·
智谱发布 GLM-5.3-FlashX：输出速度达 200 tokens/s，主打国产算力推理
智谱 AI 在 BigModel 平台正式上线 GLM-5.3-FlashX 大模型并开放 API，最大输出速度 200 tokens/s，主打智能、价格与速度三项优势，面向企业与开发者提供高吞吐、低延迟推理。文章称前代 GLM-5.3-Flash 曾以 Ox Alpha 名义在海外匿名出现，因同尺寸下智能与性价比表现获得关注、用量持续增长；为支撑需求，智谱依托 10 万卡国产芯片算力底座并加大推理优化投入，最终推出增强版 FlashX。FlashX 在保持原有智能水平与低价的同时提升生成速度，强化高并发场景的商业优势，开发者可通过官方 API 文档接入，非技术用户可在体验中心试用。
AI推荐理由 智谱在 BigModel 平台上线 GLM-5.3-FlashX，API 同步开放，最大输出速度 200 tokens/s，定位高吞吐低延迟推理；文中称前代 GLM-5.3-Flash 曾以 Ox Alpha 名义在海外匿名出现，并依托 10 万卡国产算力底座做推理优化。信息点具体（模型名、平台、速度指标、算力规模），但来源为 AIbase 转述，非官方一手发布，速度与算力数字未给出可验证的评测或基准，'智能水平与价格不变'也属厂商口径，需保留不确定性。对关注国产模型推理效率与 API 落地的开发者有参考价值。
62
正文
Zhipu Launches GLM-5.3-FlashX: Speed Reaches 200 Tokens/s, Domestic Computing Power Advances Again
AIbase基地
Published inAI News · 2 min read · Sep 18, 2026
7
Zhipu AI has officially launched the GLM-5.3-FlashX large model on its BigModel platform, with the API also going live. The model is identified as "GLM-5.3-FlashX." The new version achieves a maximum output speed of 200 tokens/s, and it focuses on the three core strengths of "intelligence, price, and speed," offering enterprises and developers high-throughput, low-latency inference services.
The previous version of this model, GLM-5.3-Flash, had once anonymously appeared overseas under the name "Ox Alpha." It quickly gained a reputation for its strong intelligence and high cost-effectiveness at the same size, and its usage volume kept increasing. To support the growing demand, Zhipu relied on a domestic chip computing power base of 100,000 units, increased investment in inference optimization, and finally launched the enhanced version FlashX.
FlashX maintains the original level of intelligence and affordable pricing while maximizing the generation speed, further strengthening its commercial advantages in high-concurrency scenarios. Developers can access it through the official API documentation, and non-technical users can also try it directly in the experience center. This upgrade marks another step forward for domestic large models in inference efficiency and popularization.
Zhipu AIGLM-5.3-FlashXBigModelLarge Model
This article is from AIbase Daily
Scan to view
Welcome to the [AI Daily] column! This is your daily guide to exploring the world of artificial intelligence. Every day, we present you with hot topics in the AI field, focusing on developers, helping you understand technical trends, and learning about innovative AI product applications.
—— Created by the AIbase Daily Team
© Copyright AIbase Base 2024, Click to View Source -