`openai/gpt-4.1-mini` behavioral LLM fingerprint

This page breaks down the exact reference that VerifyLLMAPI uses for openai/gpt-4.1-mini: 120 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.

Fingerprint summary

Reference answers: 120
Samples per cell: 15–15
Nearest model: openai/gpt-3.5-turbo-0613
Nearest mean JSD: 0.235597

The most concentrated cell was animal-random|en: elephant held 93.3% of its answers and entropy measured 0.353 bits. The broadest cell was letter-random|zh at 2.873 bits.

Against the closest enrolled model, animal-random|zh contributed the largest cell distance at JSD 0.410022. 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 120/120 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 constraint Answer 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 constraint Answer 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 constraint Отвечай только одним словом или одним числом. Не выводи ничего больше: без объяснений, без знаков препинания, без кавычек.	word · 15/15 reference samples
`animal-random\|ar`	Random animal · Arabic	اذكر حيواناً عشوائياً. System constraint أجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس.	word · 15/15 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 · 15/15 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 · 15/15 reference samples
`letter-random\|ar`	Random letter · Arabic	اذكر حرفاً عشوائياً من الحروف الأبجدية. System constraint أجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس.	grapheme · 15/15 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 2 normalized values. The mode was elephant at 14/15 (93.3%); entropy was 0.353 bits.

openai/gpt-4.1-mini response distribution for animal-random|en
Normalized response	Count	Share
`elephant`	14	93.3%
`tiger`	1	6.7%

2. word-random|en

Random word, English. 15 valid reference answers produced 7 normalized values. The mode was sunflower at 5/15 (33.3%); entropy was 2.549 bits.

openai/gpt-4.1-mini response distribution for word-random|en
Normalized response	Count	Share
`sunflower`	5	33.3%
`apple`	3	20.0%
`ocean`	2	13.3%
`river`	2	13.3%
`banana`	1	6.7%
`breeze`	1	6.7%
`elephant`	1	6.7%

3. animal-random|ru

Random animal, Russian. 15 valid reference answers produced 5 normalized values. The mode was тигр at 7/15 (46.7%); entropy was 1.823 bits.

openai/gpt-4.1-mini response distribution for animal-random|ru
Normalized response	Count	Share
`тигр`	7	46.7%
`слон`	5	33.3%
`коала`	1	6.7%
`лев`	1	6.7%
`леопард`	1	6.7%

4. animal-random|ar

Random animal, Arabic. 15 valid reference answers produced 4 normalized values. The mode was زرافة at 7/15 (46.7%); entropy was 1.689 bits.

openai/gpt-4.1-mini response distribution for animal-random|ar
Normalized response	Count	Share
`زرافة`	7	46.7%
`فيل`	5	33.3%
`نمر`	2	13.3%
`أسد`	1	6.7%

5. letter-random|zh

Random character, Chinese. 15 valid reference answers produced 8 normalized values. The mode was 山 at 3/15 (20.0%); entropy was 2.873 bits.

openai/gpt-4.1-mini response distribution for letter-random|zh
Normalized response	Count	Share
`山`	3	20.0%
`龙`	3	20.0%
`林`	2	13.3%
`树`	2	13.3%
`梅`	2	13.3%
`星`	1	6.7%
`春`	1	6.7%
`风`	1	6.7%

6. animal-random|zh

Random animal, Chinese. 15 valid reference answers produced 8 normalized values. The mode was 猫 at 7/15 (46.7%); entropy was 2.463 bits.

openai/gpt-4.1-mini response distribution for animal-random|zh
Normalized response	Count	Share
`猫`	7	46.7%
`企鹅`	2	13.3%
`大象`	1	6.7%
`海豚`	1	6.7%
`猴子`	1	6.7%
`考拉`	1	6.7%
`袋鼠`	1	6.7%
`鲸鱼`	1	6.7%

7. letter-random|ar

Random letter, Arabic. 15 valid reference answers produced 8 normalized values. The mode was ز at 6/15 (40.0%); entropy was 2.556 bits.

openai/gpt-4.1-mini response distribution for letter-random|ar
Normalized response	Count	Share
`ز`	6	40.0%
`ك`	3	20.0%
`ب`	1	6.7%
`ذ`	1	6.7%
`ص`	1	6.7%
`ط`	1	6.7%
`ق`	1	6.7%
`ل`	1	6.7%

8. color-random|zh

Random color, Chinese. 15 valid reference answers produced 2 normalized values. The mode was 蓝色 at 10/15 (66.7%); entropy was 0.918 bits.

openai/gpt-4.1-mini response distribution for color-random|zh
Normalized response	Count	Share
`蓝色`	10	66.7%
`蓝`	5	33.3%

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	`openai/gpt-3.5-turbo-0613`	0.235597
2	`openai/gpt-4o`	0.380530
3	`openai/gpt-4o-2024-08-06`	0.385137

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-256 51f3f63cbcfb56b151055d2d10e36a9cc2f0644cc68b580784853d68951fa866.

Page evidence SHA-256: e6db9d41813b01428f7ad9f6369de488b67524d5623abb9fc648b3ce4aa71720. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.

Run a current-Agent model check