Detailed Pricing

MeshAPI provides a simplified token-based pricing structure to ensure you only pay for exactly what you consume.

How Token Pricing Works

Like many foundational model APIs, usage is calculated per token. We display pricing based on blocks of 1,000 tokens.

There are two primary components to pricing:

• Prompt Pricing (prompt_usd_per_1k): The cost per 1,000 tokens of context sent to the model.
• Completion Pricing (completion_usd_per_1k): The cost per 1,000 tokens generated by the model.

Since different models require different computational resources, pricing varies heavily. Some lighter models are exceptionally cheap per token, while heavy reasoning models cost more.

Free Tier vs. Paid Models

We sort models into two distinct buckets to make cost administration easy:

1. Free Models: These models are completely free to use (is_free = true) and cost $0 for both prompt and completion. They are exceptional for testing your application routing, standardizing prompts, and light tasks.
2. Paid Models: These charge a variable rate. Paid models typically include advanced reasoning capabilities and larger context sizes.

Pre-paid Balances

Before using Paid models, you must load funds onto your account balance. Your balance is debited micro-fractions of a cent per token processed. Once your balance reaches zero, paid API requests will return an HTTP 402 “Spend Limit Exceeded” response until you top up.