Tracked providers (live)
Live provider data could not be refreshed for this build. Showing the last verified snapshot.
| Provider | Input /1M tokens | Output /1M tokens |
|---|---|---|
| Z.ai | $1.40 / 1M | $4.40 / 1M |
| OpenRouter | $1.40 / 1M | $4.40 / 1M |
| Vercel AI Gateway | $1.40 / 1M | $4.40 / 1M |
What is GLM-5.3?
GLM-5.3 is Z.ai's flagship mixture-of-experts model — ~744B total parameters with ~40B active per token, 1,048,576-token native context window, native multimodal input (text, image, video, file), text-only output, and tool calling enabled. It uses the same base model as GLM-5.2 but adds scaled-up post-training, lifting agentic-coding scores while changing output length and tool-use behavior.
Specs at a glance
| Field | Value |
|---|---|
| Vendor | Z.ai (Zhipu AI) |
| Total parameters | ~744B |
| Active parameters | ~40B per token |
| Architecture | MoE |
| Context window | 1,048,576 tokens |
| Modalities in | Text, image, video, file |
| Modalities out | Text only |
| Tool calling | Supported |
| Thinking mode | Always on, cannot be disabled |
| Recommended settings | temperature 1, top_p 0.95, reasoning_effort max, thinking_clear thinking false |
| Price, input | $1.40 per 1M tokens |
| Price, output | $4.40 per 1M tokens |
| Independent index | 60 on Artificial Analysis Intelligence Index |
Ready to test the workflow?
Create account & add creditsWhere to start
- Need API examples? — /models/zai-org/GLM-5.3/api
- Comparing providers and live prices? — /models/zai-org/GLM-5.3/providers
- Full pricing breakdown? — /models/zai-org/GLM-5.3/pricing
- Self-hosting or GPU requirements? — /models/zai-org/GLM-5.3/gpu-requirements
- Benchmark scores? — /models/zai-org/GLM-5.3/benchmarks
- Alternatives and rivals? — /models/zai-org/GLM-5.3/alternatives
Provenance
Catalog data pulled 31 August 2026 from Z.ai docs, Artificial Analysis, Hugging Face, and the OneInfer catalog entry for `z-ai/GLM-5.3`. Provider numbers are independently verifiable against the OpenRouter listing; benchmark scores are split between vendor-reported and the one independently reproduced Artificial Analysis Intelligence Index (60).
Frequently asked questions
Is GLM-5.3 multimodal?
Yes — it accepts text, image, video, and file input natively and returns text output. The flash sibling GLM-5.3-Flash is also multimodal.
How big is GLM-5.3?
~744B total parameters with ~40B active per token. Full-precision self-hosting needs on the order of 1.5TB of VRAM; aggressive 2-bit quantization brings that to roughly 245GB. Open weights were not yet released as of 31 August 2026.
How much does GLM-5.3 cost?
$1.40 per 1M input tokens, $4.40 per 1M output tokens on the Z.ai / OpenRouter listing as of late August 2026.
Where is the live GLM-5.3 model page?
The canonical model page with current OneInfer pricing, capabilities, and availability is /models/zai-org/GLM-5.3. This page is a focused facet of that entity, not a replacement for it.
How should I treat benchmark or price claims?
Check each claim’s provenance label and observed date. Vendor-reported and independently verified numbers are shown as separate evidence classes on this hub.
Put GLM-5.3 to work
Fund a controlled evaluation, start with a prepared prompt, and measure quality and cost on your own workload.