The example demonstrating `maxOutputTokens` without `thinkingBudget` was originally labeled as incorrect. This change applies the fix suggested in the comments (adding a `thinkingBudget`) and updates the section header and comments to present it as a correct example of handling low token limits.