Z.ai: GLM 5.3 FlashX
About this model
GLM-5.3-FlashX is Z.ai's high-throughput serving tier for GLM-5.3-Flash, released on September 18, 2026. It runs the same 320-billion-parameter mixture-of-experts checkpoint with roughly 18 billion active parameters per token, native multimodal inputs (text, image, video, and files), up to one million tokens of context, and always-on extended reasoning with configurable reasoning effort. FlashX is optimized for peak generation speeds of about 200 tokens per second on Z.ai's infrastructure rather than a separate weight release or benchmark harness.
Benchmark Scores
Technical Specs
- Parameters: 320.0B
- Architecture: MoE Transformer with hybrid sparse and linear attention
- Context Window: 1,000,000 tokens
- Input Modalities: text, image
Hardware Requirements
- Compute: API only
Pricing
| Input | Output | Currency |
|---|---|---|
| 0.37 / 1M tokens | 1.25 / 1M tokens | USD |
Related Models
More Multimodal →
This model always redirects to the latest model in the Claude Fable family.
This model always redirects to the latest model in the Claude Opus family.
Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...
Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...