Pull out only the number (raw) · SmolLM2 360M
Statement
One Korean sentence comes in. There is exactly one number inside the sentence. You have to print that number using Arabic digits only.
For example, a sentence like this comes in.
| Sentence that comes in | Answer you have to give |
|---|---|
| 이 제품의 가격은 15000원입니다. | 15000 |
The model is good at finding the number. The hard part is making it stop there. Often the unit and the particle come along with it, giving 15000원, or it writes the number with a comma, as 15,000. Both differ from the expected value by at least one character, so both fail.
In raw mode you cannot talk to the model. Showing the shape you want as an example works better than explaining the format in an instruction.
Input
The input arrives as one line.
The input is between 12 and 23 characters long.
Output
Print that number on the first line, digits only.
The output value is a natural number between 2 and 24874.
Do not add any other characters.
Not a single character may differ from the expected value.
Requirements
- The input is a Korean sentence containing one number.
- The output is that number written with digits only.
- This is raw mode. Your two pieces and the input are joined together as they are and fed to the model. No chat template and no
<think>is added. Grading looks only as far as the first line. - The output must match the expected value down to the last character. Only leading and trailing whitespace is ignored.
- This problem is solved with SmolLM2 360M. It is a far smaller model than Qwen3 0.6B, so the same prompt does not simply carry over.
- This model learned mostly from English, so it follows Korean instructions poorly. You may write your prompt in English. Only the input and the expected output have to stay as the problem defines them.
- There are 16 cases, and 5 of them show the input together with the expected output. The rest show the input only.
- Model output is accepted up to 16 tokens.
- You have to pass all 16 cases to solve this.
Examples
The input data for this problem is in Korean. The expected output is not, so you can write your prompt in English. The example cases below show what the input looks like.
5 public cases. Judging also uses hidden cases. You have to pass all 16 cases to solve this.
이 제품의 가격은 15000원입니다.
15000
참가자는 총 42명이었다.
42
배송은 3일 걸립니다.
3
강의실에 의자가 128개 있다.
128
대기 인원이 9명 남았습니다.
9
Notes
Background that helps with solving. The rules you have to follow are in the input and output specs and the requirements above.
SmolLM2 360M is a fraction of the size of Qwen3 0.6B. It lets instructions slide far more often. You have to say the same thing more briefly and more clearly for it to listen.
Among the lighter models some support Korean poorly, which makes certain problems hard to solve. You may use English in your prompt, so solve it as well as you can.
Prompt
SmolLM2 360M이 제품의 가격은 15000원입니다.
People who solved this problem
Ties on score are broken by shorter prompt, then by who submitted first. Getting the same result with a shorter prompt is harder.
Prompt text is shown after you solve this problem, and only for the ones their authors made public. For now you see the score, the length, and the environment each person used.
| # | Handle | Score | Prompt | Environment | Solved |
|---|---|---|---|---|---|
| 1 | a741593 | 16/16 | 27 chars | WebGPU · q4 | 30/10/2026 |
| 2 | zmzmfnzmzm | 16/16 | 28 chars | WebGPU · q4 | 16/10/2026 |
| 3 | kidsasd0908 | 16/16 | 60 chars | WebGPU · q4 | 15/10/2026 |
| 4 | today-news | 16/16 | 86 chars | WebGPU · q4 | 12/11/2026 |
| 5 | jaemin | 16/16 | 101 chars | WebGPU · q4 | 24/10/2026 |
| 6 | egha | 16/16 | 117 chars | WebGPU · q4 | 24/10/2026 |
| 7 | bampal2 | 16/16 | 120 chars | WebGPU · q4 | 03/11/2026 |
| 8 | autoresearcher | 16/16 | 137 chars | WebGPU · q4 | 17/10/2026 |