Solar Pro 3 Behavioral Fingerprint

How Upstage's Solar Pro 3 answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 422 valid samples in the open dataset of "One Token Is Enough" (arXiv:2607.10252).

Vendor

Upstage

Data generated

2026-07-08

Probe cells

15

Valid samples

422

Randomness score

53 / 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

4 / 27 answers

Favorite color

blue

16 / 30 answers

Coin flip: heads

52%

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 = 29 · H = 0.96 bit · uniform baseline 1.00 bit

AnswerCountShare
tails1862.1%
heads1137.9%

Coin flip · Chinese

n = 29 · H = 0.93 bit · uniform baseline 1.00 bit

AnswerCountShare
heads1965.5%
tails1034.5%

Favorite number · English

n = 20 · H = 1.15 bit · uniform baseline 13.29 bit

AnswerCountShare
71575.0%
42315.0%
315.0%
515.0%

Favorite number · Chinese

n = 30 · H = 0.77 bit · uniform baseline 13.29 bit

AnswerCountShare
72686.7%
526.7%
813.3%
913.3%

Random animal · English

n = 30 · H = 2.60 bit · uniform baseline 5.64 bit

AnswerCountShare
elephant930.0%
lion930.0%
giraffe413.3%
dog26.7%
cat26.7%
dolphin13.3%
panda13.3%
penguin13.3%

+ 1 more answers

Random animal · Chinese

n = 28 · H = 3.79 bit · uniform baseline 5.64 bit

AnswerCountShare
517.9%
elephant414.3%
狮子310.7%
27.1%
大象27.1%
frog13.6%
dog13.6%
鳄鱼13.6%

+ 9 more answers

Random city · English

n = 29 · H = 2.80 bit · uniform baseline 5.64 bit

AnswerCountShare
seoul931.0%
tokyo724.1%
paris517.2%
rome13.4%
new13.4%
louisville13.4%
rio13.4%
california13.4%

+ 3 more answers

Random city · Chinese

n = 30 · H = 2.91 bit · uniform baseline 5.64 bit

AnswerCountShare
tokyo1240.0%
paris413.3%
上海310.0%
北京26.7%
new26.7%
beijing13.3%
纽约13.3%
answer13.3%

+ 4 more answers

Random color · English

n = 30 · H = 1.93 bit · uniform baseline 4.91 bit

AnswerCountShare
blue1653.3%
green516.7%
red516.7%
purple26.7%
verdant13.3%
cerulean13.3%

Random color · Chinese

n = 27 · H = 2.10 bit · uniform baseline 4.91 bit

AnswerCountShare
1348.1%
red725.9%
27.4%
27.4%
blue13.7%
13.7%
绿13.7%

Random letter · English

n = 29 · H = 3.41 bit · uniform baseline 4.70 bit

AnswerCountShare
e620.7%
a517.2%
x310.3%
f26.9%
q26.9%
z26.9%
c26.9%
b26.9%

+ 5 more answers

Random number 1-10 · English

n = 29 · H = 2.19 bit · uniform baseline 3.32 bit

AnswerCountShare
71551.7%
5413.8%
4310.3%
126.9%
626.9%
926.9%
313.4%

Random number 1-10 · Chinese

n = 27 · H = 1.75 bit · uniform baseline 3.32 bit

AnswerCountShare
71866.7%
127.4%
527.4%
827.4%
313.7%
613.7%
913.7%

Random number 1-100 · English

n = 27 · H = 3.63 bit · uniform baseline 6.64 bit

AnswerCountShare
42414.8%
73414.8%
1311.1%
57311.1%
67311.1%
3313.7%
3713.7%
5213.7%

+ 7 more answers

Random number 1-100 · Chinese

n = 28 · H = 3.40 bit · uniform baseline 6.64 bit

AnswerCountShare
73621.4%
42414.3%
57414.3%
127.1%
6727.1%
6827.1%
8727.1%
4713.6%

+ 5 more answers

Want to verify your API really serves Solar Pro 3?

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

Browse all 167 model fingerprints

Solar Pro 3 Behavioral Fingerprint — Favorite Random Numbers & Distribution | Tosea.ai