GLM-5.3
Long-horizon coding on open weights at a fraction of frontier pricing.
GLM-5.3 supersedes GLM-5.2, GLM-5.1.
What are GLM-5.3's specs and price?
GLM-5.3, built by Z.ai, ships a 1M-token context window and a 131K-token max output, released 2026-08. It supports text input with a dedicated reasoning mode and costs $2.15 per million blended tokens, the 29th-cheapest of 49 models we track.
What are GLM-5.3's specs?
| Context window | 1M tokens |
| Max output | 131K tokens |
| Modalities | text |
| Extended thinking | Yes |
| Released | 2026-08 |
| Knowledge cutoff | Not published |
| Provider | Z.ai |
Verified 2026-10-01 — source.
Where does GLM-5.3 rank?
What are GLM-5.3's strengths?
- Z.ai coding-first flagship, 50% stronger than 5.2 on code bench
- 1M-token context with 128K outputs
- Mandatory reasoning with low/high/max effort
What else should you know about GLM-5.3?
What are common questions about GLM-5.3?
What is GLM-5.3's context window?
GLM-5.3 has a 1M-token context window and a 131K-token max output — the 21st-largest context of the 49 current models we track. Source: https://docs.z.ai/guides/llm/glm-5.3, verified 2026-10-01.
Does GLM-5.3 support vision or audio input?
No — GLM-5.3 is text-only as of 2026-10-01.
Does GLM-5.3 have a reasoning or extended-thinking mode?
Yes — GLM-5.3 exposes a dedicated reasoning mode for multi-step problems.
When was GLM-5.3 released, and what is its knowledge cutoff?
GLM-5.3 was released 2026-08.
How much does GLM-5.3 cost, and who provides it?
GLM-5.3 is served by Z.ai at $2.15/M blended tokens (3:1 input:output) — the 29th-cheapest of 49 current models. Full pricing breakdown: /llm-api-pricing/glm-5-3.