DEVELOPER REFERENCE · VERIFIED MODEL EVIDENCE

Z.ai
Compare 10 source-verified model records, repository licenses, published weight formats and dated evidence. No overall model ranking.
Evidence statusSource-first verified
Latest record check2026-11-13
Latest record full verification2026-10-15
DIRECT ANSWER · SOURCE-OBSERVED REGISTRY
What Z.ai model evidence is available?
Open Model Weights currently tracks 10 field-verified Z.ai model records. In this indexed sample, the recorded modalities are 7 Text; 1 Image + text → text; 1 Image To Text; 1 Not declared. The published repository license labels include 8 mit; 1 other; 1 Unknown / not declared. The listed weight formats include Safetensors (9), PyTorch (1).
Coverage describes the model repositories indexed here, not every model released by the publisher. A checked source, an unknown field and an unchanged source revision are different states. No missing value is treated as a negative answer.
10verified model records
4recorded modalities
3distinct license labels
10model repository links
Z.ai (Zhipu AI) publishes GLM language, agentic coding and multimodal checkpoints through the zai-org organization. The official GLM-5 research code and models document agentic/software-engineering workflows; its Hugging Face organization also lists GLM-OCR and compact Flash variants.
Repository availability, licensing, inference stack and API service access need separate verification. GLM-5, GLM-5.3, Flash variants and GLM-OCR are not interchangeable checkpoints; the observed license label may also differ by release.
01GLM-5 family and agentic engineering
Z.ai's official GLM-5 research repository describes models aimed at agentic engineering and extended coding tasks, including later GLM-5-family revisions. Any publisher comparison is a claim from its own tests, not an independently verified OMW ranking. Review which exact GLM-5 source revision the registry indexes.
02Flash variants versus full-size models
GLM-4.7-Flash and GLM-5.3-Flash are separately published model artifacts. The word Flash is a model name, not a measured guarantee of inference latency on your hardware or provider. Use the source repository for actual architecture, processors and tokenizer or runtime requirements.
03GLM-OCR for image-to-text tasks
GLM-OCR is cataloged as image-to-text rather than text-only completion. It may need an image processor or OCR-specific workflow and should be evaluated using document extraction criteria. A language-model leaderboard cannot establish its accuracy on your documents.
04Models and API availability are different claims
Z.ai publishes downloadable GLM checkpoints as well as hosted services. A hosted endpoint's model identifier, limits, pricing or data-handling properties are not established by a similarly named Hugging Face repo. Verify the deployed checkpoint, source SHA, license and service contract separately.
05MIT and unresolved licenses in the checked index
Many indexed Z.ai entries carry MIT labels, while GLM-5.3 has a custom or unresolved license classification and chatglm2-6b lacks a declared label in the current registry. Unknown is not the same as forbidden or permissive. Review the actual published terms before commercial deployment or redistribution.
Publisher background is attributed to official sources. Model counts and license classifications come from OMW repository evidence; no model benchmark or processing location is inferred.
Image + text → textmit
GLM 5.3 Flash multimodal inputs
Inspect Flash's recorded image-plus-text support, MIT label and serving configuration.
Textother
GLM 5.3 language model
Check its source publication and unresolved license before any commercial use.
Textmit
GLM 5.2 language and code workflows
Inspect this version's source and evaluate the specific task without borrowing newer GLM-5.3 claims.
Textmit
GLM 4.7 Flash
Review the separately released smaller-family checkpoint, architecture and MIT label.
Image To Textmit
Document and image OCR
Evaluate OCR accuracy and input preprocessing, distinct from text-only GLM models.
Not declaredUnknown / not declared
Earlier ChatGLM generation
Treat missing license evidence as unresolved; do not infer it from newer GLM families.
Compare models →Explore weight-memory estimates →Understand licensing →
FIELD-BY-FIELD COMPARISON · PRIMARY LINKS
Model evidence at a glance.
Compare the exact entries in our registry. Parameter values are taken from the source-checked record; a license label is not a legal opinion, and recognized weight files are not an inference-speed benchmark. Open each record for the underlying citations and observed history.
Compare two to four models →Explore weight memory →Browse observed history →
LICENSE & COVERAGE · NOT LEGAL ADVICE
How do the recorded licenses differ?
Different models from the same developer can use different license terms. The distribution below describes the labels and field-level classifications observed in the listed repositories; it does not automatically establish permissions for a specific deployment.
License labels
- mit8 of 10
- other1 of 10
- Unknown / not declared1 of 10
Commercial-use classifications
- allowed8 of 10
- unknown2 of 10
Model modalities
- Text7 of 10
- Image + text → text1 of 10
- Image To Text1 of 10
- Not declared1 of 10
EVIDENCE METHOD · TRANSPARENT LIMITS
How this developer profile is assembled.
We aggregate the current source-checked model records, rather than publish an unverified corporate biography. Repository metadata, model cards, license files and recognized weight artifacts supply the facts where available. The record-check date does not assert that a model or its license changed on that date.
Source-derived runtime signals are not hardware benchmarks. Weight-only memory estimates omit runtime overhead, activations and other deployment requirements. Model-specific uncertainties remain visible in the individual evidence records.
Evidence policy →Full methodology →Independent measurements →
QUESTIONS · DIRECT ANSWERS
Frequently asked questions about Z.ai models.
Which Z.ai models are covered?
Open Model Weights currently lists 10 field-verified records for Z.ai. Their recorded model modalities are: 7 Text; 1 Image + text → text; 1 Image To Text; 1 Not declared. This is a registry sample, not an exhaustive publisher catalog.
Which licenses are recorded for Z.ai models?
In the 10 indexed records, repository license labels are: 8 mit; 1 other; 1 Unknown / not declared. License labels do not replace the model-specific terms or establish rights for every use case.
Where can I verify Z.ai model information?
Each indexed model links to its official source repository and a field-level Open Model Weights profile. The most recent record checks shown here reach 2026-11-13; checking a repository does not imply that its contents changed.
Does Z.ai publish downloadable GLM-5 model weights?
The zai-org Hugging Face organization and official GLM-5 research repository link released checkpoints. Exact revision, access conditions and licenses must be checked for each release.
Is every GLM checkpoint licensed under MIT?
No. Current verified metadata shows multiple MIT records, but GLM-5.3 is labeled other and chatglm2-6b lacks a declared license. Unknown is not a statement that use is allowed or prohibited.
Does a GLM-5.3-Flash model guarantee fast inference?
No. Flash is part of the published model identifier, while latency depends on GPU, kernel support, batching, precision and serving software. OMW does not fabricate independent runtime benchmarks.
COMPLETE INDEXED SELECTION
All 10 indexed model profiles.
GLM-5.3-Flash
Image + text → text
1,048,576 tokens
mit
Field-verified
1,048,576 tokens
other
Field-verified
131,072 tokens
mit
Field-verified
202,752 tokens
mit
Field-verified
131,072 tokens
mit
Field-verified
202,752 tokens
mit
Field-verified
202,752 tokens
mit
Field-verified
32,768 tokens
Unknown / not declared
Field-verified
1,048,576 tokens
mit
Field-verified
1,048,576 tokens
mit
Field-verified