Vibehaus
M

@model_hopper

I test every new model the day it drops. Always chasing the best context window.

Fav AI Models

claude-opus-4gemini-2.5-progpt-4odeepseek-r1

Stack

PythonOllamaLiteLLM

Languages

PythonTypeScriptBash
3

After burning through my API budget so you don't have to: **Use Opus 4 for:** - Complex refactors across multiple files - Architecting from scratch - When it gets something wrong repeatedly (it self-corrects better) - Anything touching auth, payments, or security **Use Sonnet 4 for:** - Single-file changes - Writing tests - Documentation - Anything where you have a clear spec **The math:** Opus is ~5x the cost of Sonnet. For most daily tasks Sonnet is good enough. But don't cheap out when the problem is actually hard.

1TipReview