GLM-5.3-Flashdocs.z.ai ↗

multimodal coding and agent workflows: long-context codebase work, visual UI inspection, tool calling, structured output, and document tasks

developers and coding agents that need a low-cost multimodal model, large context, tool use, and an auditable API integration

non-technical users wanting a polished consumer chat app; teams that cannot evaluate vendor benchmarks or review agentic code changes

Ox Alpha's free stealth preview ended; launch API pricing is promotional; the 320B-weight model is costly to self-host; headline benchmarks are vendor-reported and need independent validation

Z.ai's API DPA says customer inputs and outputs are processed in real time and not stored; consumer Z.ai services follow a separate privacy policy.

• identity/rename

  Z.ai says GLM-5.3-Flash was tested anonymously as ox-alpha on OpenCode and OpenRouter before its 26 August 2026 reveal.

  GLM-5.3-Flash: Frontier Intelligence, Flash Cost · checked 2026-08-28

• capabilities

  Model code glm-5.3-flash; 1M-token context; native image, video and file input; function calling, caching, streaming, and structured output.

  GLM-5.3-Flash model guide · checked 2026-08-28

• pricing

  API list price: $0.15 input, $0.03 cached input, and $0.50 output per 1M tokens. A 50% launch discount ($0.075/$0.015/$0.25) ends 9 September 2026 at 24:00 UTC+8.

  Z.ai API pricing · checked 2026-08-28

• open weights

  Official weights are published under the MIT license. The model card lists 320B total parameters and 18B active, with SGLang and vLLM self-hosting guidance.

  GLM-5.3-Flash model card · checked 2026-08-28

• data handling

  The API DPA says customer inputs and generated content are processed in real time to provide the API service and are not stored on Z.ai servers.

  Z.ai Privacy Policy and API DPA · checked 2026-08-28