Text-to-SQL data analyst: what it costs and which AI to use
Staff ask questions in plain English and the model writes the SQL against your schema. Data teams and product teams giving non-analysts self-serve access to the warehouse.
The short answer
At 6,000 queries a month, we estimate $2.54 on Qwen3.8 Flash (the cheapest model that cleared our quality bar), $14.84 on Kimi K2.6, and $113 on GPT-5.5. The flagship costs 7.6x the middle pick. Judged on SQL and reasoning tasks: a query that runs but returns the wrong rows is the worst failure.
Simulate your text-to-SQL data analyst
Pick a job, set how much of it you do, and see what every model would cost. Defaults are typical, not yours. Change them.
Prompt, history, documents
What the model writes
Advanced: caching, reasoning, retries, discounts
Share that can use cached-input prices where a model offers them
1 = none. 2 = hidden thinking doubles billed output
Extra calls on top of the base count
Batch jobs or committed spend. Check your provider; we assume none by default
Lowest benchmark score a model can have and still be picked. Higher is stricter.
Limit to certain providers
Untick any provider you cannot use, for data-residency, contract or policy reasons. Picks are recalculated.
6,000 queries a month, three ways to pay
Every model here cleared a score of 9.4 out of 10 on our sql, reasoning benchmark tasks, ran without errors and fits a 16K-token window. 23 of 386 priced models qualify.
Cheapest model that clears the bar.
Qwen3.8 Flash
- Per query
- $0.0004
- Benchmark
- 10.00 / 10 · 6 prompts
- Context
- 1M tokens
Priced near the middle of the models that clear the bar; best score there.
Kimi K2.6
- Per query
- $0.0025
- Benchmark
- 10.00 / 10 · 6 prompts
- Context
- 262K tokens
What many people reach for first.
GPT-5.5
- Per query
- $0.019
- Benchmark
- 9.33 / 10 · 6 prompts
- Context
- 1.05M tokens
Choosing GPT-5.5 over Kimi K2.6 costs $97.69 more a month, $1,172 a year (7.6x the price).
On our benchmark the middle pick scored higher (10.00 against 9.33). The extra money buys safety margin our benchmark cannot measure: it is 24 short prompts, and most passing models score near the top. If a wrong answer is costly for you, that margin can be worth it. Test on your own data before you decide.
If it grows
Pay-per-use cost scales in a straight line. Monthly cost at 1x, 3x, 10x and 100x today's volume.
| Model | 1x · 6,000 | 3x · 18K | 10x · 60K | 100x · 600K |
|---|---|---|---|---|
| Qwen3.8 Flash | $2.54 | $7.62 | $25.38 | $254 |
| Kimi K2.6 | $14.84 | $44.51 | $148 | $1,484 |
| GPT-5.5 | $113 | $338 | $1,125 | $11,253 |
What a budget buys
- Qwen3.8 Flash236Kqueries a month, about 7,879 a day
- Kimi K2.640Kqueries a month, about 1,347 a day
- GPT-5.55,331queries a month, about 177 a day
Every model, priced for your workload
41 models, 11 hidden because their window is smaller than 16K tokens.
| Model | A month | Per query | Benchmark | Context |
|---|---|---|---|---|
| DeepSeek V4 Flash 0423 DeepSeekbelow the bar | $1.45 | $0.0002 | 8.67 (6) | 1.05M |
| DeepSeek V4.1 Flash DeepSeekbelow the bar | $1.66 | $0.0003 | 9.33 (6) | 1.05M |
| Ministral 3 8B 2512 Mistral AIbelow the bar | $1.89 | $0.0003 | 4.50 (6) | 262K |
| Llama 4 Scout Metabelow the bar | $2.28 | $0.0004 | 7.00 (6) | 1.31M |
| Qwen3.8 Flash AlibabaBudget pick | $2.54 | $0.0004 | 10.00 (6) | 1M |
| Mistral Small 4 Mistral AI | $2.78 | $0.0005 | 9.83 (6) | 262K |
| GLM 5.3 Flash Z Aibelow the bar | $2.79 | $0.0005 | 8.50 (6) | 1.05M |
| GLM 5.3 Z Ai | $4.37 | $0.0007 | 9.83 (6) | 1.05M |
| GPT-5.4 Nano OpenAI | $4.60 | $0.0008 | 9.83 (6) | 400K |
| MiniMax M2.5 Minimax | $4.79 | $0.0008 | 10.00 (6) | 205K |
| Qwen3.8 27B Alibababelow the bar | $5.62 | $0.0009 | 8.33 (6) | 1.02M |
| Muse Glimmer 30B Meta | $5.70 | $0.0010 | 9.50 (6) | 131K |
Benchmark = mean judge score out of 10 on the sql, reasoning prompts of the Model Playground, with the number of prompts in brackets. Models marked n/a have no benchmark run, so we do not guess their quality. How the benchmark works.
What we assumed
- Six thousand questions a month.
- Each question sends the relevant schema and a few example queries, about 3,500 tokens.
- The SQL comes back at about 300 tokens.
- 60% of input (the schema) repeats and can be cached. 10% of queries are retried after an error.
These are typical numbers, not measurements of your business. Edit them in the simulator above.
What drives the cost
The database schema rides along with every question, so input is large and the output is a short query.
Watch out for
- Warehouse compute for running the generated queries is not included. It is often bigger than the model bill.
- Add a human or automated check before generated SQL touches production data.
Questions people ask
How much will text-to-SQL data analyst cost per month?
At 6,000 queries a month, we estimate $2.54 a month on Qwen3.8 Flash, the cheapest model that clears our quality bar for this job. Kimi K2.6 costs $14.84 and GPT-5.5 costs $113. These are estimates from the assumptions listed on this page, not a quote. Change the numbers to match your own volume.
What is the cheapest AI model for text-to-SQL data analyst?
On current OpenRouter prices, Qwen3.8 Flash from Alibaba is the cheapest model that cleared our bar of 9.4 out of 10, scoring 10.0/10 on 6 benchmark prompts for this kind of work. The benchmark is small, so treat it as a shortlist and test on your own data before switching.
Is it worth paying for a flagship model for text-to-SQL data analyst?
On our benchmark, GPT-5.5 scored 9.3/10 on this kind of work, and a model priced near the middle of the models that cleared the bar, Kimi K2.6, scored 10.0/10, while the flagship costs 7.6x as much. A flagship buys safety margin the benchmark cannot measure. If a wrong answer is expensive for you, that can be worth paying for. If not, you are probably overpaying.
What happens to the cost if text-to-SQL data analyst grows ten times?
Pay-per-use pricing scales in a straight line, so ten times the queries is ten times the bill. Use the growth table on this page to see 3x, 10x and 100x for the same three models.
What makes text-to-SQL data analyst expensive?
The database schema rides along with every question, so input is large and the output is a short query.
Price a different job
- Customer support chatbotFrom $1.74 a month at 3,000 conversations
- Coding assistant for a teamFrom $4.12 a month at 4,400 requests
- Document summarisationFrom $1.86 a month at 2,000 documents
- Blog and content writingFrom $0.083 a month at 120 articles
- Knowledge base Q&A (RAG)From $5.62 a month at 15K questions
- Invoice and form data extractionFrom $4.65 a month at 20K documents
- Ticket and content classificationFrom $11.30 a month at 500K items
- AI agent workflowsFrom $7.44 a month at 1,500 runs
- Sales email personalisationFrom $0.595 a month at 5,000 emails
- Meeting notes and summariesFrom $0.575 a month at 400 meetings
- Code review for pull requestsFrom $0.840 a month at 600 pull requests
- Personal AI assistant for a founderFrom $0.478 a month at 1,500 messages