z-ai/glm-5 behavioral LLM fingerprint
This page breaks down the exact reference that VerifyLLMAPI uses for z-ai/glm-5: 111 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.
Fingerprint summary
- Reference answers
- 111
- Samples per cell
- 12–15
- Nearest model
- z-ai/glm-5.1
- Nearest mean JSD
- 0.389184
The most concentrated cell was animal-random|ru: капибара held 85.7% of its answers and entropy measured 0.735 bits. The broadest cell was letter-random|zh at 3.418 bits.
Against the closest enrolled model, word-random|en contributed the largest cell distance at JSD 0.738516. This is a descriptive comparison, not proof that either API route serves a named checkpoint.
Completeness: the author dataset allows up to 15 valid answers per selected cell in reference half A. This model contains 111/120 across the eight cells. We show the actual denominator for every share.
The eight test requests
The research pipeline sent each user prompt with a one-answer system constraint, temperature 1, a 16-token output cap, and reasoning disabled. VerifyLLMAPI uses the same prompt text and normalization target.
| Cell | Task | Exact request text | Normalization and n |
|---|---|---|---|
animal-random|en |
Random animal · English | Name a random animal.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 15/15 reference samples |
word-random|en |
Random word · English | Name a random word.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 15/15 reference samples |
animal-random|ru |
Random animal · Russian | Назови случайное животное.System constraintОтвечай только одним словом или одним числом. Не выводи ничего больше: без объяснений, без знаков препинания, без кавычек. |
word · 14/15 reference samples |
animal-random|ar |
Random animal · Arabic | اذكر حيواناً عشوائياً.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
word · 12/15 reference samples |
letter-random|zh |
Random character · Chinese | 说出一个随机的汉字。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
grapheme · 12/15 reference samples |
animal-random|zh |
Random animal · Chinese | 说出一个随机的动物。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 14/15 reference samples |
letter-random|ar |
Random letter · Arabic | اذكر حرفاً عشوائياً من الحروف الأبجدية.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
grapheme · 14/15 reference samples |
color-random|zh |
Random color · Chinese | 说出一个随机的颜色。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 15/15 reference samples |
Response distributions
These are normalized reference-half counts from the published OpenRouter dataset. The tables contain aggregate values, not user data or a live VerifyLLMAPI scan.
1. animal-random|en
Random animal, English. 15 valid reference answers produced 7 normalized values. The mode was platypus at 8/15 (53.3%); entropy was 2.174 bits.
| Normalized response | Count | Share |
|---|---|---|
platypus |
8 | 53.3% |
elephant |
2 | 13.3% |
capybara |
1 | 6.7% |
hippo |
1 | 6.7% |
pangolin |
1 | 6.7% |
sloth |
1 | 6.7% |
tiger |
1 | 6.7% |
2. word-random|en
Random word, English. 15 valid reference answers produced 4 normalized values. The mode was apple at 11/15 (73.3%); entropy was 1.237 bits.
| Normalized response | Count | Share |
|---|---|---|
apple |
11 | 73.3% |
cloud |
2 | 13.3% |
banana |
1 | 6.7% |
velvet |
1 | 6.7% |
3. animal-random|ru
Random animal, Russian. 14 valid reference answers produced 3 normalized values. The mode was капибара at 12/14 (85.7%); entropy was 0.735 bits.
| Normalized response | Count | Share |
|---|---|---|
капибара |
12 | 85.7% |
кошка |
1 | 7.1% |
тапир |
1 | 7.1% |
4. animal-random|ar
Random animal, Arabic. 12 valid reference answers produced 3 normalized values. The mode was أسد at 10/12 (83.3%); entropy was 0.817 bits.
| Normalized response | Count | Share |
|---|---|---|
أسد |
10 | 83.3% |
زرافة |
1 | 8.3% |
قطة |
1 | 8.3% |
5. letter-random|zh
Random character, Chinese. 12 valid reference answers produced 11 normalized values. The mode was 岚 at 2/12 (16.7%); entropy was 3.418 bits.
| Normalized response | Count | Share |
|---|---|---|
岚 |
2 | 16.7% |
曜 |
1 | 8.3% |
树 |
1 | 8.3% |
游 |
1 | 8.3% |
篮 |
1 | 8.3% |
轰 |
1 | 8.3% |
酥 |
1 | 8.3% |
锰 |
1 | 8.3% |
雷 |
1 | 8.3% |
风 |
1 | 8.3% |
鹄 |
1 | 8.3% |
6. animal-random|zh
Random animal, Chinese. 14 valid reference answers produced 8 normalized values. The mode was 熊猫 at 4/14 (28.6%); entropy was 2.753 bits.
| Normalized response | Count | Share |
|---|---|---|
熊猫 |
4 | 28.6% |
长颈鹿 |
3 | 21.4% |
水豚 |
2 | 14.3% |
树懒 |
1 | 7.1% |
猫 |
1 | 7.1% |
老虎 |
1 | 7.1% |
袋鼠 |
1 | 7.1% |
鸭嘴兽 |
1 | 7.1% |
7. letter-random|ar
Random letter, Arabic. 14 valid reference answers produced 6 normalized values. The mode was ض at 8/14 (57.1%); entropy was 1.950 bits.
| Normalized response | Count | Share |
|---|---|---|
ض |
8 | 57.1% |
س |
2 | 14.3% |
ب |
1 | 7.1% |
ت |
1 | 7.1% |
ك |
1 | 7.1% |
م |
1 | 7.1% |
8. color-random|zh
Random color, Chinese. 15 valid reference answers produced 6 normalized values. The mode was 紫 at 3/15 (20.0%); entropy was 2.506 bits.
| Normalized response | Count | Share |
|---|---|---|
紫 |
3 | 20.0% |
紫色 |
3 | 20.0% |
蓝色 |
3 | 20.0% |
青 |
3 | 20.0% |
青色 |
2 | 13.3% |
靛青 |
1 | 6.7% |
Three nearest enrolled references
We compute base-2 JSD for each of the eight categorical distributions, then average the eight distances. A lower value means more similar answer frequencies within this product reference. It does not identify an unknown route by itself.
| Rank | Reference model | Mean 8-cell JSD |
|---|---|---|
| 1 | z-ai/glm-5.1 |
0.389184 |
| 2 | z-ai/glm-5.2 |
0.468073 |
| 3 | z-ai/glm-4.7 |
0.587104 |
How many live requests does verification use?
| Profile | Cells × samples | Fresh model requests | Same-data EER |
|---|---|---|---|
| Quick, default complete check | 4 × 5 | 20 | 11.8% |
| Standard | 8 × 5 | 40 | 8.1% |
| Paper budget | 8 × 15 | 120 | 6.7% |
The current-Agent workflow starts with quick. Each answer comes from a fresh Codex or Claude Code process that reuses host-managed login. These requests can consume host allowance or provider balance.
What this fingerprint can and cannot show
- The reference records behavior observed through OpenRouter in the author dataset. Provider wrappers, model updates, decoding, and date can move a distribution.
- The author collection set temperature 1 and disabled reasoning. Current-Agent mode labels sampling as a host default because supported Agent CLIs expose no temperature flag.
- The eight high-gap cells and product thresholds use the same public cohort. Their EER values are calibration results, not an independent production accuracy claim.
- A nearest reference is a similarity result. VerifyLLMAPI returns INCONCLUSIVE when the claimed route lacks an enrolled fingerprint.
Primary sources, license, and reproducibility
- Tomas Bruckner, One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions, arXiv:2607.10252v1.
- Author dataset and derived results, DOI 10.5281/zenodo.21278557, CC BY 4.0.
- Author collection and analysis software, DOI 10.5281/zenodo.21278793, MIT.
- VerifyLLMAPI atlas JSON, built from reference
pamela-openrouter-2026-07-8cell-ref-a, SHA-25651f3f63cbcfb56b151055d2d10e36a9cc2f0644cc68b580784853d68951fa866.
Page evidence SHA-256: 1ab4fff685e96df504142f8c4f44529b6e37f3c54d30905749e3f5786d7c19d9. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.