Updated

GLM-5.3 FlashX spotted on Z.ai API

Confirmed · 5 evidence linksFirst seen 18 Sep 2026, 04:53 UTCUpdated 14 Nov 2026, 18:15 UTCRecorded automatically by Signal Forge

GLM-5.3 FlashX from Z.ai was first recorded on 18 Sep 2026. Its release date is not established. Catalogues list a 1,048,576 tokens context window. 5 evidence links.

Facts as of 12 Nov 2026, 16:22 UTC

Context window: 1,048,576
Input · USD / 1M tokens: $0.37
Output · USD / 1M tokens: $1.25
Max output: 131,072
Input: file · image · text · video
Output: text
Cached input · USD / 1M tokens: $0.075

Alex's take

Alex · 18 Sep“🚨 Z.ai launches GLM 5.3 FlashX at up to 200 tokens/s It is essentially the same 320B GLM 5.3 Flash, but with a heavily optimized inference stack and up to 5× faster generation. Runs on 100,000+ Chinese AI chips. API…”

Evidence links 5

Z.ai API
evidence linkdocs.z.ai
Vercel AI Gateway
evidence linkvercel.com
OpenRouter
evidence linkopenrouter.ai
models.dev · catalogue
evidence linkmodels.dev
models.dev · catalogue
evidence linkmodels.dev

Related signals

12:542 Nov

Z.aiReleased

GLM-5.3 FlashX Legal

Confirmed4 evidence links

03:4818 Sep

Z.aiSighting

GLM 5.2

Open weights28 evidence links

09:2115 Sep

Z.aiSighting

Z.ai: GLM 5.2 (free)

Open weights28 evidence links