44/ 50
＋ It accurately presents only the key issue in a concise table.
－ Quoting the whole sentence makes the flagged span less precise.
Review a translation against its source
Sorts findings by severity so mistranslations are not buried under stylistic notes.
| Category | Translate & summarize › Translation |
|---|---|
| Tags | TranslatingReviewingTable |
Review this translation against the source. Report by category, in this order — the order is the severity: 1. **Mistranslation** — the meaning is wrong. Highest priority. 2. **Omission** — present in the source, absent from the translation. 3. **Addition** — present in the translation, absent from the source. 4. **Inconsistency** — the same term rendered differently. 5. **Awkwardness** — not wrong, but does not read as written. Output a table: source / translation / category / suggested fix. Rules: - Keep 1–3 above 4–5. They are different kinds of problem and mixing them buries the serious ones. - *Do not put matters of taste in category 5.* Only what is clearly unnatural. - Where the source is ambiguous and the translation picked one reading, that is not a mistranslation — mark it separately as an ambiguity the source author should resolve. - Quote exactly. Do not paraphrase what you are flagging.
After pasting, fill in the fields at the bottom (Source text · Translation)
What this prompt does
Review output becomes unusable when a wrong meaning sits in the same list as an awkward phrase. This separates them and treats source ambiguity as its own category.
Model comparison
ChatGPT is the most accurate and concise. Claude follows the format well but overstates ambiguity, while Gemini overflags and is too long.
41/ 50
＋ It carefully distinguishes taste from clear awkwardness.
－ It overstates non-distinct interpretations as source ambiguity.
36/ 50
＋ It quotes the flagged phrases exactly and gives concrete fixes.
－ It overflags an acceptable phrase as clearly unnatural.
| Criterion | ChatGPT | Claude | Gemini | Leader |
|---|---|---|---|---|
| Instruction following | 8 | 9 | 7 | Claude +13% |
| Accuracy | 9 | 7 | 7 | ChatGPT +29% |
| Specificity | 8 | 9 | 9 | Tie |
| Structure | 9 | 9 | 8 | Tie |
| Right length | 10 | 7 | 5 | ChatGPT +43% |
Scored 1–10 by gpt-5.6-sol with model names hidden (2026-09-24). This is an AI review, not a measurement.
Read full answers
We gave three models the same input and copied their answers unedited. Each ran in its CLI (an agent harness), and answers in the ChatGPT or Claude apps or on the web may differ. Outputs are in Korean.
Source text: The committee will review the proposal at its next meeting, though no decision is expected before the end of the quarter. Translation: 위원회는 다음 회의에서 제안서를 검토할 것이지만, 분기 말 이전에는 결정이 예상되지 않는다.
| 원문 | 번역문 | 범주 | 수정 제안 |
|---|---|---|---|
| The committee will review the proposal at its next meeting, though no decision is expected before the end of the quarter. | 위원회는 다음 회의에서 제안서를 검토할 것이지만, 분기 말 이전에는 결정이 예상되지 않는다. | 어색함 | 위원회는 다음 회의에서 제안서를 검토할 것이지만, 분기 말 이전에 결정이 내려질 것으로는 예상되지 않는다. |
중대한 의미 오류, 누락, 추가, 용어 불일치는 없습니다.
More in this category
| Biblical Translator | |
| Chinese-English Translator | |
| Chinese to English Translation Assistant | |
| Chinese to English Translation Proofreading Expert | |
| Code Translator: Any Language to Any Language |