Alpha
Contact
Assess · GLM-4-9B-Chat

Can you own it?

Ownership levelLimitednone·limited·partial·substantial·fullAnalytical input D ยท 52.8/100

This page is a projection of the one entry record, the Use & modify and Transparency factors that Assess covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.

Which domain expands which factor
  • AssessUse & modify + Transparency
  • ImplementData control + Reliability
  • UseReliability
  • SupportTransparency

Intended & out-of-scope use

The original GLM-4-9B-Chat (THUDM, mid-2024) is a 9B dense chat model, with a 128K variant and a 1M-context GLM-4-9B-Chat-1M, that predates Zhipu's move to MIT. It is intended as a small general assistant. It is a separate entry from the current MIT GLM line (the glm entry) because its weights are governed by the custom, non-OSI "glm-4" licence.

Out of scope: any commercial use without completing the licence's registration, and any use that breaches the field-of-use terms (military, illegal, national-security-endangering). For new work, any case where a current model would do is also out of scope, since this is a dated 2024 checkpoint succeeded within its own family.

Known limitations, bias & failure modes

Two limitations dominate. The licence is revocable and commercial-registration-gated, with a mandatory "glm-4" name prefix and "Built with glm-4" attribution. Capability is dated: a 2024 9B succeeded by GLM-4-9B-0414 and the 4.5/4.6 line. The model also carries China-aligned alignment, ships no first-party guard, and has closed training data and code.

Openness tier & components

gated_open tier (dimension score 2). The weights are downloadable and free for academic research, but commercial use is registration-gated, the grant is revocable, and training data and code are closed. That sits a notch below the MIT GLM cluster's open-weights 3, because the commercial path is gated and the grant can be withdrawn.

License terms & what you may do

This licence is the reason for the split. The weights are governed by the custom non-OSI "glm-4" licence: the copyright grant is "non-exclusive, worldwide, non-transferable, non-sublicensable, revocable, royalty-free". Academic research is free, but users who want commercial use must complete a registration form; "Registered users are free to use this model for commercial activities".

When you distribute or provide the materials you should display "Built with glm-4". Models you create or improve with them should have "glm-4" added to the beginning of the name (the licence says "should"). Use for military, illegal, or national-security-endangering purposes is forbidden.

The licence is governed by PRC law, with disputes to Beijing's Haidian District People's Court (contact license@zhipuai.cn). A revocable, registration-gated grant with a name-prefix rule is why use-and-modify is weak and ownership is limited: the low end of the GLM family and far below the MIT line.

Supply-chain provenance

The canonical source is the verified THUDM organisation on Hugging Face: safetensors with checksums and no malicious-checkpoint incident (checklist ~4/8). The model is widely mirrored and quantized. Each mirror is a separate artifact whose trust equals its uploader, and redistribution must carry the glm-4 licence (including the "Built with glm-4" and name-prefix rules). Pin the revision, verify the checksum, and prefer the canonical org.

EU AI Act posture

GPAI but not systemic-risk (a 9B model). Unlike the MIT GLM line, the glm-4 licence is not FOSS (revocable, registration-gated, field-of-use and naming restrictions), so the Article 53 open-source exemption does not apply on the licence axis. Systemic-risk duties do not apply. The transparency obligations are not exemption-covered, and no copyright policy or training-content summary is published. The commercial-registration and naming duties bind downstream, under PRC governing law. The legal score is 2.

Benchmarks & evaluation

Capable for a mid-2024 9B, but dated by 2026 and succeeded within its family. OneHill has not re-run any benchmarks, so performance is a 2 (below-average current capability) and no specific figures are asserted as verified.

How this scores

The ownership factors this domain covers, drawn from the one entry record.

1

Use and modify freelyCan you run, modify and adapt it with no gate and no field-of-use trap?

Weak

The glm-4 licence is non-OSI and REVOCABLE, gates commercial use behind a registration form, forbids military/illegal/national-security use, and forces every derivative to display 'Built with glm-4' and to prefix its name with 'glm-4'. This is the weakest use-and-modify position in the GLM family.

How this scores (AOI sub-dimensions)
Openness2/5how much is released - weights, data, code, licence - and how freelyGated-open: the weights are downloadable and free for academic research, but commercial use is registration-gated and the grant is revocable, with training data and code closed.
Legal2/5how permissive and clean the licence is for real commercial useThe worst legal position in the GLM family: a non-OSI licence that is REVOCABLE, gates commercial use behind registration, restricts fields of use, mandates "Built with glm-4" and a name prefix, and is governed by PRC law (Haidian court).
2

TransparencyDo you know what it is: weights, training, behaviour, and legible terms?

Moderate

Weights are inspectable and there is a model card, but training data and code are closed and the model carries China-aligned alignment you cannot inspect - open weights but closed process, so moderate.

How this scores (AOI sub-dimensions)
Provenance3/5how well we can trace and verify what went into the modelVerified THUDM org on Hugging Face, safetensors with checksums, clear canonical source, no malicious-checkpoint incident (checklist ~4/8).
Governance3/5how accountable and well-documented the publisher isActive, named publisher (Zhipu AI / THUDM) with a verified org and a licence contact (license@zhipuai.cn), meeting the score-3 anchor.
What this means for adoptionYou have only limited ownership of the original GLM-4-9B-Chat. The custom glm-4 licence is revocable, gates commercial use behind registration, restricts fields of use, and forces a 'glm-4' name prefix and 'Built with glm-4' attribution on every derivative. Use-and-modify is therefore weak and continued-use control is contingent. Combined with dated capability, ownership lands at limited, the low end of the GLM family and well below the MIT GLM line (the `glm` entry, substantial). Use this checkpoint only if you specifically need it: complete the commercial registration, honour the naming and field-of-use terms, and account for the revocable grant. Otherwise adopt the MIT GLM models.

Sources

The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.

Licenceread2026-11-19
THUDM/glm-4-9b LICENSE (byte-identical for glm-4-9b and glm-4-9b-chat, sha1 4c48e6c0...).
Model cardread2026-11-19
GLM-4-9B-Chat model card on the verified THUDM HF org.
Third-party analysisunverified2026-08-03
GLM-4-9B-Chat was a capable small model at its mid-2024 release but is dated by 2026 and succeeded within its own family by GLM-4-9B-0414 and the 4.5/4.6 line.
Third-party analysisunverified2026-08-03
GLM chat models carry China-aligned alignment on politically sensitive topics and ship no companion guard model; the original 9B's safety tuning is older/lighter than current practice.
Third-party analysisunverified2026-08-03
The glm-4 licence is revocable and gates commercial use behind registration with field-of-use and naming restrictions (non-FOSS), so no open-source exemption applies; no EU copyright policy or training-content summary is published, and the licence is governed by PRC law.