GLM-5.3-FlashX
The higher-throughput variant of GLM-5.3-Flash, provisioned for peak traffic and burst workloads. It trades a little unit price for faster served tokens and steadier latency under load.
At a glance
Released
Sep 18, 2026
Provider
Zhipu AI
Context window
128,000 tokens
Max output
32,000 tokens
Modalities
Text → text
Knowledge cutoff
June 2026
API model ID
glm-5.3-flashxLicense
Open weights (MIT)
Pricing
Standard
Input$0.37 / 1M
Output$1.25 / 1M
A typical 1,500-token prompt with a 650-token answer costs ≈ $0.001 on the standard tier. Run your own prompt through the calculator →
What's new vs the previous model
Compared withGLM-4.7-FlashX (May 2026)
- Higher sustained throughput than GLM-4.7-FlashX
- Improved instruction adherence over the prior FlashX
- Priced above the plain Flash tier
Best use cases
High-traffic endpoints
Keeps latency flat when request volume spikes.
Batch pipelines
Chews through large jobs faster than the base Flash tier.
Real-time features
Powers autocomplete and inline assists that need snappy responses.
Strengths & limitations
Strengths
- High throughput
- Stable under load
- Open weights
Limitations
- Costs more than base Flash
- Same reasoning ceiling as Flash
How to test it
- No-code: open chat.z.ai and pick GLM-5.3-FlashX from the model switcher — no API key needed.
- Playground: sign in to the Z.ai API console and run prompts in the workbench, then copy the generated code.
- API: create a key in the Z.ai API console and call the API (example below).
curl https://api.z.ai/api/paas/v4/chat/completions \
# OpenAI-compatible endpoint
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "glm-5.3-flashx",
"messages": [{"role": "user", "content": "Explain tokenization in one paragraph."}]
}'