Release tracker / Z.ai
Z.ai model releases
Every Z.ai model launch we have recorded, newest first: GLM small, GLM. Each one links to its source, and to how it ranks on our benchmarks where we have tested it.
- Latest
- GLM-5.3-Flash, 20 Aug 2026
- Launches
- 10
- Since
- 17 Jan 2024
- Since last launch
- 53 days
Z.ai's latest model
-
$0.15 in · $0.50 out per MEpoch AI source
GLM-5.3-Flash came out 53 days ago. Spotted a launch we're missing? Tell us.
How often Z.ai ships
Lines with at least three launches in the last two years. Amber: longer since the last launch than the line's average gap. Grey: no launch for 90 days or more, marked quiet.
| Line | Size | Launches | Average gap | Latest | Last launch |
|---|---|---|---|---|---|
| GLM | Large | 7 | 63 days | GLM-5.314 Aug 2026 | 59 days ago |
Every Z.ai release
-
$0.15 in · $0.50 out per MEpoch AI source
-
#16 of 22 overall$1.40 in · $4.40 out per MReference source
-
$1.40 in · $4.40 out per MReference source
-
$1.38 in · $4.40 out per MReference source
-
$0.95 in · $2.55 out per MReference source
-
$0.54 in · $1.98 out per MReference source
-
$0.50 in · $2.00 out per MReference source
-
$0.60 in · $2.20 out per MReference source
-
$0.14 in · $0.86 out per MEpoch AI source
Questions
- What is Z.ai's latest model?
- GLM-5.3-Flash, released 20 Aug 2026 (53 days ago). It is the newest of the 10 Z.ai launches we have recorded since 17 Jan 2024.
- When will Z.ai release its next model?
- Z.ai has not announced its next model. It usually releases a new GLM model about every 63 days; the last was GLM-5.3, 59 days ago.
- How often does Z.ai release new models?
- Over the last two years: a new GLM model about every 63 days (7 launches).
Other providers: OpenAI, Mistral AI, Alibaba, Anthropic, Google, DeepSeek, xAI, Meta, Moonshot AI, Microsoft, MiniMax, Xiaomi, Amazon, NVIDIA, StepFun. Or see every launch on one timeline.
A launch is the first day a model was publicly available, previews included; a family's sizes released together count once, and later dated updates of a model are left out. Main lines are checked by hand against the provider's own changelog or announcement where we could read it ("Provider"), otherwise a reference page ("Reference"); the rest come from Epoch AI's model data (CC BY 4.0).