A company needs multilingual product summaries within a defined latency and cost budget. Several supported models are available. What is the most appropriate selection process?

1. Use an English-only demonstration and assume every target language will perform identically.
2. Evaluate candidate models on representative languages and products, measuring summary quality, latency, supported limits, and total inference cost.
3. Select the cheapest input-token price alone and ignore output cost and answer quality.
4. Select the largest model automatically and ignore latency and unit-cost measurements.