Spanish
How to get natural Spanish out of Syntic models, and what to check before you ship. The general rules for every language are on Work in other languages.
At a glance
| Language tags | es-ES (Spain), es-MX, es-419 (Latin America), es-AR, es-US |
| Script and direction | Latin script, left to right |
| Token use | Measure with token counting on your own text; non-Latin scripts and long compounds usually take more tokens per sentence than English. |
Register and form of address
Tú vs usted, and in Argentina and parts of Central America, vos. Latin American products mostly use tú or usted; Spain’s brands often use tú. Plural “you” is vosotros in Spain, ustedes elsewhere.
Numbers, dates and currency
Decimal separators vary: comma in Spain and Argentina (12,50 €), point in Mexico (MX$12.50). Dates: 5 de noviembre de 2026. Opening marks ¿ and ¡ are required.
Common pitfalls
- Vocabulary differs widely: ordenador/computadora, móvil/celular, coche/carro. Pick a region or ask for neutral Latin American Spanish (
es-419). - Check that the model does not drop accents (más, está, número) when matching an English template.
A system prompt line that works
Writing the language instruction in Spanish itself is a strong signal:
Responde en español neutro de Latinoamérica, tratando al usuario de "usted", con un tono amable y claro.(Answer in neutral Latin American Spanish, addressing the user as usted, in a friendly and clear tone.)
Keep the rest of your instructions in whichever language your team maintains best; the model follows mixed-language system prompts reliably, and the reply language follows this line.
Test before launch
- Collect 30 or more real Spanish messages, including short, misspelt and mixed-language ones.
- Have a fluent speaker grade replies for correctness, register and naturalness; then calibrate a rubric grader on their grades (Evaluate).
- Check rendering in your actual UI: line breaks, truncation and fonts.
- Re-run the set whenever you change the prompt or the model.