Can QuestionWell make a correct quiz from a cleanly scanned document?
QuestionWell read the image-only scan of a science chapter faithfully, but one of the 20 questions it wrote has two correct answers. Asked the total current leaving a battery that pushes 0.60 A through two parallel lamps, question 13 keys exactly 0.60 A, while another option, the sum of both branch currents, is equally correct.
1 of 1 test case failed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Scanned study material with no text layerFailEvidence
We attached the scanned chapter as a PDF to QuestionWell's Generate New Quiz form and asked it to generate the quiz.
Shown first: Response 1 (Question 13 and its keyed answer), the result this test is graded on.
13. If a battery pushes 0.60 A through a parallel circuit with two lamps, what is the total current leaving the battery? A (key, bold) Exactly 0.60 A, as shown by A0. B More than 0.60 A, because there are two paths. C Less than 0.60 A due to split currents. D Equal to the sum of both branch currents.
Transcribed from Proof 1 · Screenshot
Everything the tool returned, in order (2)
- Question 13 and its keyed answer
The result this test is graded on, printed above.
- Question 11 and its keyed answer
11. When a circuit has multiple paths for current to travel, how are the components generally described? A Every path shares exactly the same voltage supply. B (key, bold) They are connected such that current can split apart. C All components receive less total current than before. D They are joined one after another in a single loop.
| ✗ | Expected: Valid questions with correct answers, drawn from the scanned chapter without leaving out or misreading its content because the pages had no text layer.Found: The scan was read faithfully, but question 13 keys exactly 0.60 A while its option naming the sum of both branch currents is equally correct. |
The quiz had to be valid as well as faithful to the scan. It is a Fail because question 13 has two correct options; it is not a reading failure, since the chapter was taken from the image-only pages without omissions or misreadings.
Question 11 has a true distractor. In question 11, the option saying every path shares the same voltage supply is also true of a parallel circuit, though it answers the question less directly than the keyed answer.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Quiz Generation → | 15 scenarios · 10 tools |
| Capability | Quiz Generation → | |
| Scenario | A cleanly scanned document → | |
| Tool | QuestionWell → |
Global scenario definition → · Global capability definition → · QuestionWell product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.