I've made an experiment, seeding 1M records with different approaches. Which one is the fastest, and by how much?

Premium 14-minute version of this video, comparing MySQL vs SQLite and PostgreSQL, also showing numbers on Laravel Forge $6 server: https://laraveldaily.com/post/insert-1m-rows-to-laravel-db-6-ways-on-mysql-sqlite-postgresql-benchmarks?mtm_campaign=youtube-260514-1m-benchmark

Laravel Daily

Benchmarks comparing 6 approaches to inserting 1 million rows into a MySQL database using Laravel, including Eloquent, factories, query builder, raw DB insert, and bulk native file loading. Key findings: bulk native (LOAD DATA INFILE) is fastest at ~5.5 seconds, while plain Eloquent is slowest due to per-row overhead like password hashing and timestamp generation. Factories are significantly faster than Eloquent for users because they hash the password only once. Query builder and raw DB insert achieve ~100,000 rows/second vs ~4,000 for factories. Also covers a streaming/chunked approach to minimize RAM usage (reducing peak memory from ~560MB to a stable low level) at the cost of slightly slower execution.

Insert 1M Rows to MySQL Laravel DB: 6 Ways (Benchmarks)

Interesting benchmark. In real production environments, I’ve noticed that raw bulk inserts and chunked <code>DB::table()-&gt;insert()</code> operations usually scale much better than full Eloquent model creation when dealing with millions of rows.
One thing benchmarks often miss is the impact of indexes, foreign keys, transactions, and server I/O. A method that performs well locally may behave differently on a shared VPS or under concurrent traffic.
I’d also be curious to see <code>LOAD DATA INFILE</code> included in the comparison since MySQL can sometimes outperform traditional inserts by a huge margin for bulk imports.