01 · The idea
What stayed with us
Small public SQL models provide realistic baselines and expose how much benchmark, model size, and serving format affect apparently similar claims.
02 · The local translation
What we did with it
PostTrainLLM compared cssupport/t5-small-awesome-text-to-sql and reviewed prem-1B-SQL, Qwen SQL variants, and SQLCoder-class models.
03 · The boundary
Where the comparison stops
Model-card scores from different datasets are directional context, not a head-to-head result against the local specialist.