Skip to main content
Credits are Voxcode’s unit of AI usage. Every request you make consumes credits based on the model you choose and the number of tokens processed. 10 credits ≈ $1 USD

​
How credits are calculated

Credits are consumed based on the number of input tokens (your message, code context, conversation history) and output tokens (the model’s response) in each request.
Costs are rounded to four decimal places. A typical short coding question costs a fraction of a credit; a long multi-file refactor with a premium model may cost a few tenths of a credit.
BYOK (Bring Your Own Key) models do not consume Voxcode credits. When you use BYOK, your provider bills you directly for token usage.

​
Model pricing

Voxcode groups models into three tiers. The tier a model belongs to also determines which plan gives you access to it.

​
Free models (!)

Free models cost zero Voxcode credits regardless of how many tokens you use. These are fully capable models — a great starting point for most tasks. All Free models are marked with (!) in the model picker.

​
Beta models (*)

Beta models are experimental or cutting-edge releases. Most have very low credit costs, and some are free. Beta models are marked with () in the model picker. Access requires the Starter or Premium plan.*

​
Premium models ($)

Premium models are the most capable models available in Voxcode, suited for complex, multi-step, or high-stakes tasks. Premium models are marked with ($) in the model picker. The Starter plan includes access to 2 Premium models. The Premium plan unlocks all of them.

​
Token estimation guide

Not sure how many tokens your request will use? Use these rough estimates as a guide. Most everyday coding requests (a question plus a short code response) fall in the 500–2,000 token range total.

​
Estimated requests per plan

The table below shows how many typical requests (1K input + 500 output tokens) you can make with each plan’s monthly credit allowance.

​
Best practices

  • Start with free models. Free and beta models handle the vast majority of coding tasks well. Use them to build your workflow before reaching for premium options.
  • Use fast models for simple tasks. Quick questions, syntax lookups, and one-liner fixes don’t need a top-tier model. Grok 4.1 Fast or Gemini 2.5 Flash are fast and cost almost nothing.
  • Reserve premium models for complex work. Multi-file refactors, architectural decisions, and difficult debugging sessions are where Claude Opus or GPT 5.1 Codex justify their higher cost.
  • Consider BYOK for heavy usage. If you regularly hit your credit limit, the Premium plan’s BYOK support lets you use your own API keys with no Voxcode credit consumption. See the BYOK guide for setup instructions.
You can monitor your current credit balance at any time in Settings → Usage. Top-up credits are available from your avatar → Manage Subscription whenever you need more.