What makes this Mistral Large 2 prompt unique?

This prompt has been engineered using Bayesian optimization to specifically improve Write SQL query performance. It achieves a 0% reduction in token usage compared to naive prompts.

Does this prompt work with other models besides Mistral Large 2?

While highly optimized for Mistral Large 2, the underlying structural patterns (identity enforcement, task delimitation) are highly effective across most frontier models including GPT-4 and Llama 3.

How do I use this optimized template?

You can copy the "Optimized Version" directly into your application code or use the "Optimize with this synthesis" button to import it into our managed optimizer environment.

Back to Library

Prompt Engineering Guide

Mastering Write SQL query
on Mistral Large 2

Stop guessing. See how professional prompt engineering transforms Mistral Large 2's output for specific technical tasks.

The "Vibe" Prompt

"Write a SQL query to get the top 10 customers by total order amount this year."

Low specificity, inconsistent output

Optimized Version

STABLE

You are a highly skilled SQL query generator for a PostgreSQL database. **Instructions:** 1. Generate a SQL query to retrieve the top 10 customer IDs and their total order amounts. 2. Consider only orders placed in the current calendar year (2023). 3. Assume tables `customers` (columns: `customer_id`, `customer_name`) and `orders` (columns: `order_id`, `customer_id`, `order_date`, `total_amount`). 4. The output should be a single, executable SQL query. 5. Order the results in descending order of total amount. **Thought Process:** * I need to join `customers` and `orders` tables. * I need to filter orders by `order_date` for the current year. * I need to group by `customer_id` and sum `total_amount`. * I need to order by the summed amount descending. * Finally, I need to limit the results to 10. **SQL Query:**

Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt provides a clear persona, detailed instructions, explicit table and column names, and a chain-of-thought breakdown. This specificity reduces ambiguity, guides the model through the logical steps, and helps it generate a more accurate and robust SQL query. The 'Thought Process' section acts as a strong few-shot example or a clear internal monologue for the model, improving output quality without necessarily increasing prompt length significantly for complex queries.

Token Efficiency Gain

The generated SQL uses `JOIN` correctly.

The generated SQL filters by `EXTRACT(YEAR FROM order_date) = 2023` (or equivalent).

The generated SQL uses `GROUP BY customer_id`.

How We Validate This Prompt

Every optimized prompt for Write SQL query on Mistral Large 2 is scored against a fixed set of evaluation assertions. A revision ships only when it passes all of them, so the 0% token reduction never comes at the cost of output quality.

The generated SQL uses `JOIN` correctly.
The generated SQL filters by `EXTRACT(YEAR FROM order_date) = 2023` (or equivalent).
The generated SQL uses `GROUP BY customer_id`.
The generated SQL uses `SUM(total_amount)`.
The generated SQL uses `ORDER BY SUM(total_amount) DESC`.
The generated SQL uses `LIMIT 10`.
The generated SQL selects `customer_id` and the summed amount.

Related Optimizations

Mistral Large 2Summarize documentView guide Mistral Large 2Write emailView guide Mistral Large 2Debug codeView guide Mistral Large 2Creative writingView guide GPT-4oWrite SQL queryView guide Claude 3.5 SonnetWrite SQL queryView guide

Optimize your own Write SQL query prompt

Run any prompt through the same optimizer that produced this Mistral Large 2 guide — clarity, structure, and token efficiency in one pass.

Open the Optimizer

Mastering Write SQL query on Mistral Large 2