Models · GPT
GPT-5.4
Asked for a few hundred random integers, GPT-5.4 does not pick evenly. This page is the shape of that preference, built from 36 recorded answers containing 10743 numbers, and it is what the identify tool compares a pasted reply against.
- Bank id
gpt-5.4- Family
- GPT
- Collected through
- API
- Status
- Inherited from ModelTrace; serving identity not independently verified
- Distinct values used
- 355 of 355
- Prompt environments
- 12 of 12
Distribution
Every bar is one value from 1 to 355; its height is how many times the model chose it across all recorded answers. A perfectly random picker would give a flat line. Nothing in the bank does.
Favourite numbers
| Number | Times chosen | Share |
|---|---|---|
| 58 | 42 | 0.39% |
| 54 | 41 | 0.38% |
| 287 | 41 | 0.38% |
| 27 | 40 | 0.37% |
| 57 | 40 | 0.37% |
| 84 | 40 | 0.37% |
| 97 | 40 | 0.37% |
| 12 | 39 | 0.36% |
| 17 | 39 | 0.36% |
| 63 | 39 | 0.36% |
| 24 | 38 | 0.35% |
| 118 | 38 | 0.35% |
How reliably it is recognised
In held-out testing across all twelve prompt environments, 9 of 12 three-answer runs named this model, 2 narrowed it to its family, 1 gave no answer, and 0 named a different model.
With this model removed from the bank and its own answers scored against the rest, 0 of 12 runs still confidently named another model, 7 stopped at the family, and 5 gave no answer. Its answers landed most often on GPT-6 Astra (5 of 12). That is the number to keep in mind if you are testing a version that is not listed here.
Both measurements use the open-set verdict the site applies, not the bank's raw ranking. The methodology page explains how they were made and what they leave out.