There is a conversation that has been going on in backend development circles for over a decade now, and it refuses to die. PHP or Node.js? Which ...
For further actions, you may consider blocking this person and/or reporting abuse
When you are using AI always check the facts. The versions places your post in 2024.
In 2025 PHP 8.5 and Node 24 came out.
Respect the veteran take, David, but this was a deliberate choice. Node 22 is the current LTS and 8.4 is the production standard for stability right now. Chasing the newest minor version for a benchmark gives 'lab' results, not real-world ones. This is about what we’re actually deploying in 2026, not just what's on the download page.
Node 24 is the latest LTS version.
PHP is on a one year cycle. The latest PHP version is the stable version.
So no they are not minor versions.
You're right on the release cycles—Node 24 is the active LTS now, and PHP 8.5 is the stable standard. I’ll take that hit on the version numbers but however the core of our argument stands: whether it's Node 22 or 24, the event loop still chokes on heavy CPU tasks, and whether it's PHP 8.4 or 8.5, the shift toward persistent workers (FrankenPHP) is where the real performance evolution is happening. I benched the "battle-tested" versions most conservative teams are still actually running, but I hear you—staying current is non-negotiable. I'll sharpen the focus for the next run.
I have no comments on the the rest of your post.
You framed the post in 2026, and that made the versions out of date.
If you framed the post as production tested, the versions would not have bothered me. Node 22 and PHP 8.4 have still enough lifetime to use in 2026.
That is why I thought the versions where picked by an LLM, because they have a cut off point.
I’d rather bench what’s currently powering 90% of active production apps in 2026 (22 LTS and 8.4) than the 'newest' versions that teams are still scared to migrate to. But you’re right—if I frame it as the absolute cutting edge of 2026, I have to include Node 24 and PHP 8.5.
Appreciate the call-out. It keeps the discourse honest.
Which do you think is better ?
Neither is "better"—it's all about the use case! PHP for speed of dev, Node for real-time. Pick the tool that fits the job.
Great breakdown of FrankenPHP! It’s definitely a game-changer for modern PHP performance. It really blurs the lines between these two runtimes in 2026.
Thanks, Vinod! FrankenPHP really is the "secret sauce" for PHP in 2026. Moving away from the traditional "spin up/tear down" model of FPM and keeping workers persistent brings PHP so much closer to Node's throughput. It’s an exciting time to be a PHP dev!
The CPU-bound benchmark is a huge eye-opener. Most people focus strictly on I/O, but seeing how PHP handles Fibonacci vs Node really highlights the thread model.
Spot on, Faraz! That’s the "hidden" detail that often gets lost in the hype. While Node.js is a beast for I/O, the single-threaded event loop can become a literal bottleneck the moment you throw heavy math at it. It’s all about understanding the trade-offs of the architecture you're choosing!
Got thrill reading this
Glad you enjoyed the deep dive, Omar. Thanks for the support!
Can you write on Tailwind & Bootstrap ? Would love to read that too
I’ll definitely add a Tailwind vs. Bootstrap comparison to my list. Stay tuned!
Solid comparison. It’s refreshing to see a neutral take that respects both ecosystems instead of the usual "X is dead" narrative. Great work on this.
I really appreciate that, Aley. The "X is dead" trope is exhausting and usually just clickbait. Both ecosystems are thriving and borrowing the best ideas from each other. I’m glad the neutral perspective resonated with you!
Your Origin Story was very unique 👌
Thanks, Zohaib! Glad you liked the history lesson. Understanding where they started explains a lot about how they work today.
Love the "Database is the bottleneck" point. It’s easy to get lost in micro-benchmarks, but real-world latency almost always comes down to the query.
Exactly, Faique. We spend hours arguing over milliseconds in runtime execution while a single unoptimized JOIN or a missing index costs us seconds. In the real world, the "winner" is usually the dev who optimizes their queries, regardless of the language!
Appreciate the nuance here. Most people just shout "X is better" without looking at the DB bottlenecks.
Exactly! It’s the Achilles' heel that catch people off guard once they hit production scale.
It was a Solid comparison and the CPU-bound benchmark is a huge eye-opener for me.
Thanks, Sagar! I’m glad that part was helpful. It’s definitely the one test that forces a rethink of how we distribute tasks between the main thread and background workers. Thanks for reading!
Can you make some post on Golang vs Rust ?
Go vs Rust is definitely on my list. Both are beasts for systems and high-scale backends. I'll get to it soon.
It was such a detailed but Abit outdated like acc to 2024 not 2026 current one where PHP introduced several new stuff but all overall so much detailed, impressive work Hats off
Fair point. I focused on stability over the latest RCs, but I hear you. Glad you found the depth useful regardless.