Channels & Pricing

A channel represents a source of model resources available to the DDTokens backend. Different channels may come from different providers, so their prices, speed, capacity, and parameter compatibility may differ.

When Channels Matter

Most users can use a default model name directly, for example:

gpt-4o-mini

You only need to consider channels when:

• You need to pin a particular provider.
• A model parameter behaves differently across providers.
• You want to compare price and speed.
• Request logs show an unexpected processing channel.

Default Routing

With default routing, DDTokens selects an available channel. This suits general applications because you do not need to maintain a channel list and can get better availability.

Select a Channel

If the platform supports provider prefixes, use a format like:

openai/gpt-4o-mini
anthropic/claude-3-5-sonnet
google/gemini-1.5-flash

Available prefixes, models, and prices are determined by the dashboard or official pricing page.

Pricing Notes

The same model may have different prices on different channels. Before launch, use real requests to verify:

• The model is available.
• Parameters are compatible.
• Streaming works correctly.
• Actual costs match expectations.

For more information, see Providers & Routing and Model Pricing.