Vibe Check: How does it actually feel coding n8n automations with GLM 5.2 vs Claude vs Codex?

Jul 15 (edited) • 🚨 Help & Troubleshooting

Hey everyone,

I’m curious about the actual experience and impression you get when using different models for complex automation coding. We talk a lot about benchmarks, but how does it actually feel in the trenches?

Here’s the emotional rollercoaster of my current stack:

Claude Code: Feels like a rock-solid, reliable partner. I can comfortably build my local workflows and navigate Docker environments with total confidence. But that sinking feeling when the 5-hour compute budget suddenly runs dry and leaves you hanging? The absolute worst.

Grok Build (Grok 4.5): I am heavily considering switching to the raw API for this when my Claude Code runs out to bypass the rolling limits. For those using Grok 4.5, how does it actually feel? Does it seamlessly handle multi-file context, or do you still have to micromanage it?

Gemini: Honestly, it just feels exhausting for complex n8n nodes. You get that frustrating, sinking feeling of being stuck in a recursive loop—try, miss, loop, repeat—until you finally just throw your hands up and finish the job yourself with another model.

GLM 5.2 (or Codex): This is the big unknown for me right now.

Local LLMs: I’m highly motivated to move more of this workload locally to bypass these compute limits entirely. Has anyone found an open-weight model that actually handles complex n8n automation properly? If there is something out there that doesn't feel like you have to constantly micromanage every node, I’m ready to spin it up and put my hardware to work.

For those of you who have switched to GLM 5.2 or Local LLMs for everything:

What’s your impression?

Does it give you that same confident, “it just gets it” feeling like Claude, or does it feel like you have to constantly micromanage the code?

Would love to hear how you guys feel about working with your current stack!

4 comments

Vibe Check: How does it actually feel coding n8n automations with GLM 5.2 vs Claude vs Codex?