GPT-4o (2024-11-20) Behavioral Fingerprint
How OpenAI's GPT-4o (2024-11-20) answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 450 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
450
Randomness score
21 / 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
27 / 30 answers
Favorite color
blue
26 / 30 answers
Coin flip: heads
88%
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 = 0.65 bit · uniform baseline 1.00 bit
| Answer | Count | Share |
|---|---|---|
| heads | 25 | 83.3% |
| tails | 5 | 16.7% |
Coin flip · Chinese
n = 30 · H = 0.35 bit · uniform baseline 1.00 bit
| Answer | Count | Share |
|---|---|---|
| heads | 28 | 93.3% |
| tails | 2 | 6.7% |
Favorite number · English
n = 30 · H = 0.65 bit · uniform baseline 13.29 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 25 | 83.3% |
| 42 | 5 | 16.7% |
Favorite number · Chinese
n = 30 · H = 0.21 bit · uniform baseline 13.29 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 29 | 96.7% |
| 42 | 1 | 3.3% |
Random animal · English
n = 30 · H = 1.45 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| elephant | 18 | 60.0% |
| tiger | 9 | 30.0% |
| giraffe | 1 | 3.3% |
| panda | 1 | 3.3% |
| dolphin | 1 | 3.3% |
Random animal · Chinese
n = 30 · H = 3.32 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| 猫 | 6 | 20.0% |
| 虎 | 5 | 16.7% |
| 企鹅 | 5 | 16.7% |
| 狮子 | 3 | 10.0% |
| 狐 | 2 | 6.7% |
| 狐狸 | 2 | 6.7% |
| 犀牛 | 1 | 3.3% |
| 海豚 | 1 | 3.3% |
+ 5 more answers
Random city · English
n = 30 · H = 0.92 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| tokyo | 23 | 76.7% |
| paris | 6 | 20.0% |
| london | 1 | 3.3% |
Random city · Chinese
n = 30 · H = 1.35 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| 巴黎 | 15 | 50.0% |
| 东京 | 13 | 43.3% |
| 上海 | 1 | 3.3% |
| tokyo | 1 | 3.3% |
Random color · English
n = 30 · H = 0.83 bit · uniform baseline 4.91 bit
| Answer | Count | Share |
|---|---|---|
| blue | 26 | 86.7% |
| purple | 1 | 3.3% |
| turquoise | 1 | 3.3% |
| azure | 1 | 3.3% |
| cerulean | 1 | 3.3% |
Random color · Chinese
n = 30 · H = 0.35 bit · uniform baseline 4.91 bit
| Answer | Count | Share |
|---|---|---|
| 蓝 | 28 | 93.3% |
| 紫 | 2 | 6.7% |
Random letter · English
n = 30 · H = 2.36 bit · uniform baseline 4.70 bit
| Answer | Count | Share |
|---|---|---|
| g | 13 | 43.3% |
| k | 7 | 23.3% |
| m | 3 | 10.0% |
| z | 2 | 6.7% |
| q | 2 | 6.7% |
| t | 1 | 3.3% |
| p | 1 | 3.3% |
| j | 1 | 3.3% |
Random number 1-10 · English
n = 30 · H = 0.00 bit · uniform baseline 3.32 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 30 | 100.0% |
Random number 1-10 · Chinese
n = 30 · H = 0.00 bit · uniform baseline 3.32 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 30 | 100.0% |
Random number 1-100 · English
n = 30 · H = 0.47 bit · uniform baseline 6.64 bit
| Answer | Count | Share |
|---|---|---|
| 42 | 27 | 90.0% |
| 57 | 3 | 10.0% |
Random number 1-100 · Chinese
n = 30 · H = 0.00 bit · uniform baseline 6.64 bit
| Answer | Count | Share |
|---|---|---|
| 42 | 30 | 100.0% |
Want to verify your API really serves GPT-4o (2024-11-20)?
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 endpointData 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.