Edit: this is just an out-of-the-box comparison, not a fully tuned benchmark. So please consider it before snapping out in the comments. "Pick up the fastest backend." The thing most founders say the moment you ask them about their preferred tech stack. But it's never only about speed, that decides whether your application survives in production. It's consistency under load. To prove my point in simplest possible way, I simulated 300 concurrent users on a simple blog API in two different backends (Node (Express) and FastAPI). Same database. Same endpoints. No caching. No auth. At first, both looked fine. But under load, latency spiked from 2-7 seconds in FastAPI. Node.js (Express) stayed stable with low latency and normal responses. Keep in mind, the users never care about the average 22ms result you retained while creating. They only remember the 2 second delay. #expressjs #nodejs #fastapi #backenddevelopment
What do you prefer, Speed or consistency under load?
It’s not about which is faster or better. It’s about which works better for YOU. FastAPI is naturally faster for CPU-heavy tasks like data processing and manipulation. Express is faster for I/O heavy tasks — raw requests per second. In your example, you demonstrated requests per second and that is where express scales better than FastAPI. If you tried a test with data processing FastAPI would outperform Express. It’s all about your use case — choosing the right tools for your products.
I dont think you understand how the big picture work. But sure, let's blame frameworks.
These comparisons are somewhat pointless tbh. I'll bring Go into the mix and it'll embarrass the both of these. Choose what you and you're team is good at so you can focus on your game.
trust me bro benchmarks
The comparison is okay for that schenario but not accurate. Cause you need to select the stack based on necessary. Comparing is pointless. Both has its own power for separate platform
system: python multi threading joined party system: js workers left party
What about Bun + Elysia?
As you mentioned you have database layer in both and it means you compared database driver implementation with default configuration. Where are all the details? What size of database connections pool, what query do you make etc? You compared not frameworks, but two unknown database drivers with unknown config. This is absolutely pointless to post stuff like that until you suppose to get attention from non-tech people
Do you think framework choice matters more than architecture when it comes to performance?