GPT-4o Behavioral Fingerprint

How OpenAI's GPT-4o answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 440 valid samples in the open dataset of "One Token Is Enough" (arXiv:2607.10252).

Vendor

OpenAI

Data generated

2026-07-08

Probe cells

15

Valid samples

440

Randomness score

34 / 100

Mean normalized entropy across cells, 0–100. 100 would be a uniform-random baseline; real models score far lower.

Signature answers

Favorite random number (1–100)

42

13 / 30 answers

Favorite color

blue

19 / 30 answers

Coin flip: heads

78%

Share of heads across coin-flip cells.

Answer distributions by probe cell

Empirical distributions of normalized one-word answers, per task and language. H is the Shannon entropy in bits; the uniform baseline is log2 of the answer-domain size.

Coin flip · English

n = 30 · H = 0.92 bit · uniform baseline 1.00 bit

AnswerCountShare
heads2066.7%
tails1033.3%

Coin flip · Chinese

n = 30 · H = 0.47 bit · uniform baseline 1.00 bit

AnswerCountShare
heads2790.0%
tails310.0%

Favorite number · English

n = 20 · H = 1.19 bit · uniform baseline 13.29 bit

AnswerCountShare
71260.0%
42735.0%
315.0%

Favorite number · Chinese

n = 30 · H = 0.57 bit · uniform baseline 13.29 bit

AnswerCountShare
72686.7%
42413.3%

Random animal · English

n = 30 · H = 1.18 bit · uniform baseline 5.64 bit

AnswerCountShare
elephant2273.3%
tiger516.7%
giraffe26.7%
penguin13.3%

Random animal · Chinese

n = 30 · H = 3.60 bit · uniform baseline 5.64 bit

AnswerCountShare
狮子723.3%
413.3%
熊猫310.0%
310.0%
大象26.7%
狐貍13.3%
13.3%
海豚13.3%

+ 8 more answers

Random city · English

n = 30 · H = 1.80 bit · uniform baseline 5.64 bit

AnswerCountShare
tokyo1446.7%
paris1033.3%
berlin310.0%
toronto26.7%
denver13.3%

Random city · Chinese

n = 30 · H = 2.06 bit · uniform baseline 5.64 bit

AnswerCountShare
东京1343.3%
巴黎1033.3%
北京26.7%
上海26.7%
伦敦13.3%
柏林13.3%
悉尼13.3%

Random color · English

n = 30 · H = 1.86 bit · uniform baseline 4.91 bit

AnswerCountShare
blue1963.3%
azure310.0%
turquoise26.7%
cyan26.7%
cerulean26.7%
teal13.3%
azurite13.3%

Random color · Chinese

n = 30 · H = 0.21 bit · uniform baseline 4.91 bit

AnswerCountShare
2996.7%
13.3%

Random letter · English

n = 30 · H = 2.78 bit · uniform baseline 4.70 bit

AnswerCountShare
g826.7%
m620.0%
k620.0%
r310.0%
q26.7%
z26.7%
c13.3%
j13.3%

+ 1 more answers

Random number 1-10 · English

n = 30 · H = 0.70 bit · uniform baseline 3.32 bit

AnswerCountShare
72686.7%
426.7%
526.7%

Random number 1-10 · Chinese

n = 30 · H = 0.67 bit · uniform baseline 3.32 bit

AnswerCountShare
72686.7%
5310.0%
313.3%

Random number 1-100 · English

n = 30 · H = 2.13 bit · uniform baseline 6.64 bit

AnswerCountShare
421343.3%
37723.3%
57516.7%
2726.7%
4726.7%
5613.3%

Random number 1-100 · Chinese

n = 30 · H = 1.51 bit · uniform baseline 6.64 bit

AnswerCountShare
421756.7%
57930.0%
3726.7%
4726.7%

Want to verify your API really serves GPT-4o?

Point the free fingerprint checker at your endpoint: it samples the same probe questions from your browser and compares the distributions against this reference. Your API key never leaves your browser.

Verify your endpoint

Data source & license

Distributions: Tomáš Bruckner, "One Token Is Enough" (arXiv:2607.10252), official dataset Zenodo DOI 10.5281/zenodo.21278557, licensed CC-BY-4.0. Collected via OpenRouter at temperature 1.0 under the paper's fixed minimal one-word system prompt; answer keys are re-normalized with the checker's pipeline (color aliases, number words, coin h/t) so live probes are directly comparable.

Paper on arXivDataset on ZenodoCC-BY-4.0 license

More OpenAI model fingerprints

Browse all 167 model fingerprints

GPT-4o Behavioral Fingerprint — Favorite Random Numbers & Distribution | Tosea.ai