Zukko Blog
Guides to Claude Code, Cursor, Codex and the Claude and GPT APIs: setup, token savings and troubleshooting.
Gateway or direct API access: which to choose
3 min · · pricing
Each option has its own price and its own risks. Both sides honestly, including where a gateway is inconvenient.
Grok 4.6 through the API: setup and pricing
3 min · · models
Access through a shared key, a flat rate, a million-token window. Setup and what the model is worth comparing on.
How to choose a model gateway: what to look at
3 min · · pricing
Six checks before you trust an intermediary with real traffic. Each takes a couple of minutes.
Where Qwen models fit in development work
3 min · · models
Qwen cost one and a half to three times less than flagships. Which tasks show no difference, and which do.
GPT-6 Astra through the API: setup and cost
3 min · · models
Setup takes one line. Here is the cost and a way to judge the model on your own work rather than on reviews.
Astra, Grok and Claude: how to choose between them
3 min · · models
Comparing what can actually be verified: price, window, rate shape. Plus a method for choosing under your own load.
Why your token count is higher than your prompt
2 min · · tokens
A five-word question shows up as thousands of tokens in your usage. That is not a billing bug — here is the breakdown.
Which model to choose when they all share one key
3 min · · models
When every model sits behind one key, the choice stops being technical and becomes economic. How to pick without overpaying.
What the Claude API really costs
3 min · · pricing
A rate per million tokens means nothing until you know your own workload. Here is how to actually estimate it.
Hitting Claude usage limits: what to do
3 min · · limits
The quota runs out mid-week and money cannot buy your way past it. Why the cap exists and what the alternatives actually cost.