dealignai/GLM-5.3-CYBERSECURITY-FP8
dealignai/GLM-5.3-CYBERSECURITY-FP8 is a dealignai record on this register. The Hugging Face model API lists the pipeline tag as text-generation. The safetensors index gives an exact parameter count of 753,329,940,480. Architectural context length (max_position_embeddings) is 1,048,576. The listed licence is mit. The Hub repository was created on 30 Aug 2026, which is the release date this register uses.
Primary source: the Hugging Face model card for dealignai/GLM-5.3-CYBERSECURITY-FP8. Fields below come from that card, the Hub model API, and the repository’s own config.json when those files exist. A field we cannot source is omitted rather than guessed.
| Organisation | dealignai |
|---|---|
| ReleasedRepo creation date on Hugging Face | 30 Aug 2026 |
| Last modified | 8 Sept 2026 |
| ParametersExact, from the safetensors index | 753.3B 753,329,940,480 |
| Context lengthmax_position_embeddings — architectural maximum, not necessarily served | 1M 1,048,576 |
| Licence | mit |
| Pipeline / taskpipeline_tag on the Hugging Face model API | text-generation |
| Base modelDeclared on the Hub and present on this register | zai-org/GLM-5.3 |
| File formats | safetensors, FP8 |
| Architecture | glm_moe_dsa |
| Layers | 78 |
| Vocabulary | 154,880 |
| DownloadsAt last check | 99.9k |
| Likes | 648 |
| Gated | No |
Lineage
This record names zai-org/GLM-5.3 as its base model, and that release is on this register.
How to run
Safetensors weights are on the Hugging Face repository. The parameter count on this page came from that index.
The repository name marks this record as FP8.
Load it from huggingface.co/dealignai/GLM-5.3-CYBERSECURITY-FP8. This page does not add install steps that the Hub files do not state.
Agent capabilities
Derived from the model’s own chat template, with the match counts behind each finding.
- ●Tool calling30 matches in template
- ●Reasoning / thinking15 matches in template
- ●System prompt1 match in template
- ●Image input2 matches in template
- ●Video input1 match in template
- ●Audio input1 match in template
Questions this record can answer
- When was GLM-5.3-CYBERSECURITY-FP8 released?
- The Hugging Face repository was created on 30 Aug 2026, which is the date this register uses.
- How many parameters does GLM-5.3-CYBERSECURITY-FP8 have?
- The safetensors index gives 753,329,940,480 parameters, shown here as 753.3B. That figure is exact, not rounded.
- What is the context length of GLM-5.3-CYBERSECURITY-FP8?
- 1,048,576 tokens. That is max_position_embeddings from the model's own config.json — the architectural maximum, not necessarily the context a provider serves.
- What licence is GLM-5.3-CYBERSECURITY-FP8 under?
- The Hub card lists mit.
- Is GLM-5.3-CYBERSECURITY-FP8 gated?
- No. The Hugging Face model API does not mark this repository as gated.
- Does the GLM-5.3-CYBERSECURITY-FP8 chat template show tool calling?
- The chat template contains evidence of tool calling.
Where each field came from
- Hugging Face model API — releasedAt, lastModified, downloads, likes, task, license, gated, parameters. Checked 9 Oct 2026.
- config.json — contextLength, layers, vocabSize, architecture. Checked 9 Oct 2026.
- chat template — capabilities. Checked 9 Oct 2026.
First recorded 2 Oct 2026, last checked 9 Oct 2026. This record is compiled automatically from the sources above. Where a field is blank, we could not source it and have not guessed. Canonical repository: huggingface.co/dealignai/GLM-5.3-CYBERSECURITY-FP8. This page is a quantisation or conversion. It is kept for completeness, excluded from the sitemap, and marked noindex.