Standard vs PTC: same task, which saves more
Experiment design set, results await raw records
This page gives only the experiment design, not the results. Numbers we haven’t run are never written; paste raw records into the table after running, then write the conclusion.
Question to answer
Same task, same model, same input: run both Standard mode and PTC mode once each - how much do tokens, elapsed time, and turns actually differ?
Control groups
Standard mode / PTC mode; all other conditions identical, 3 runs each, keeping raw records without averaging or smoothing.
Environment & version
- Host version
- To fill
- Model ID
- To fill
- Mode / permissions
- Keep both groups identical
- Input files
- Record hashes so both runs use the same files
Fields to log
| # | Field | Result |
|---|---|---|
| 1 | Host version and model ID | to be tested |
| 2 | Task and input-file hash | to be tested |
| 3 | Tool calls per turn | to be tested |
| 4 | Input / output tokens | to be tested |
| 5 | Cache hits and cost | to be tested |
| 6 | Wall-clock time | to be tested |
| 7 | Whether the final output matches | to be tested |
| 8 | Failure and retry records | to be tested |
Conclusion
To be tested. Any conclusion must also state what it can't generalize to.
Why we don’t write results yet
The community already reports conflicting conclusions: PTC is cheaper for repeated batch actions, but more expensive for small tasks with few files. This site does not presuppose a conclusion and will not publish numbers that have not been measured.