Teamday

Coding Model · Z.ai

GLM-5.3 Flash

GLM-5.3 Flash is Z.ai's efficient tier under GLM-5.3, released August 26, 2026 under the MIT licence: a 320B-parameter hybrid linear/sparse-attention model with 18B active parameters, native vision (text, image, and video input), and a 1M-token context window — the first natively multimodal GLM-5 model. Teamday runs it through the Pi harness via OpenRouter as the budget option below GLM-5.3.

Use GLM-5.3 Flash in Teamday →All models

At a glance (verified 2026-10-05)

Vendor
Z.ai
Type
Coding Model
Context window
1M tokens
Pricing
Z.ai's own OpenRouter endpoint lists $0.15 input / $0.03 cached / $0.50 output per 1M tokens (October 5, 2026; the 50% launch promotion has ended).
OpenRouter ID
z-ai/glm-5.3-flash

Best for

High-volume codingLong-horizon agent tasksScreenshot and UI understandingLong-context (1M) work

How to use GLM-5.3 Flash in Teamday

Pick GLM-5.3 Flash on Teamday's Pi harness, routed through OpenRouter (OPENROUTER_API_KEY).

Harness: pi →Max →

Frequently asked questions

How much does GLM-5.3 Flash cost?

Z.ai's own OpenRouter endpoint listed $0.15 per 1M input tokens and $0.50 per 1M output tokens on October 5, 2026 — the earlier 50% launch promotion has ended, though DeepInfra's route ran its own 50% promotion at $0.075 / $0.25. That is about a ninth of GLM-5.3's $1.40 / $4.40 on Z.ai's endpoint. It was the most-used model on OpenRouter by weekly tokens in mid-September 2026.

What is the difference between GLM-5.3 and GLM-5.3 Flash?

GLM-5.3 is the flagship reasoning-capable coding model. GLM-5.3 Flash is a leaner 320B / 18B-active hybrid-attention model that adds native vision, ships under the MIT licence, and costs a fraction as much — Z.ai positions it for efficient coding and long-horizon agent tasks.

Does GLM-5.3 Flash support vision?

Yes — it is the first natively multimodal model in the GLM-5 series, accepting text, image, and video input so it can observe interfaces and rendered results.

How do I run GLM-5.3 Flash in Teamday?

Select it on the Pi harness in the chat model picker, routed via OpenRouter with your own API key.

Related

GLM-5.3 Flash is one of the models Teamday's AI employees already run — no API keys, no infrastructure.

Hire your first AI employee →