Skip to main content
Channel groups are the foundation of how NexLLM manages model access and usage pricing. When you create an API key, you assign it to a channel group — that group determines which models you can call and what multiplier is applied to your base usage costs. Choosing the right group lets you balance model availability, pricing, and connectivity quality for your use case.

​
What Is a Channel Group?

A channel group is a model access and pricing category. Every API key belongs to exactly one group, and that group defines two things: the set of models you can call and the group ratio applied to the model’s base cost. The group ratio is a multiplier. A ratio below 1.0x gives you discounted pricing; a ratio of 1.0x means you pay the standard base price. If you request a model that isn’t included in your API key’s group, the request will fail — the model simply isn’t accessible through that key.

​
Price Calculation with Group Ratio

The group ratio is applied after the model’s base usage cost is calculated:
For example, if the base cost of a request is $1.00: To use discounted pricing, choose a group with a lower ratio when creating your API key — as long as that group includes the models you need.

​
Key Rules

  • Each group has its own ratio applied to the model’s base price to calculate the final usage cost.
  • Each API key belongs to exactly one group. The assigned group determines which models the key can access.
  • You can only call models included in your group. Requests to models outside your group will not be routed.
  • Lower group ratios mean discounted pricing. Select a group with a lower ratio to reduce usage costs.
  • The default group includes all models. An API key in the default group can call any available model on the platform.

​
Available Channel Groups

The group information below is provided as a reference and is subject to change. Always verify your group’s supported models before building against a specific endpoint.

​
Group Descriptions

default — Includes every available model on the platform. Use this group when you need broad access across providers and model families without restriction. claude — Provides access to Anthropic models at a 0.5x ratio, making it the most cost-effective option for Claude usage. Connectivity is acceptable, so it suits workloads that can tolerate occasional latency variation. gpt-azure — Routes OpenAI model requests through Azure at a 0.8x ratio. A good choice if you want OpenAI-compatible models with discounted pricing and reliable Azure-backed routing. gemini-vertex — Delivers Gemini text models and text-to-image generation through Vertex AI at a 0.8x ratio. Use this group when you need Gemini’s image generation capabilities alongside text models. claude-aws — Routes Anthropic model requests through AWS Bedrock at a 1.0x ratio. This group prioritizes stability and great connectivity over price discounts, making it ideal for production workloads that require consistent Claude access. gemini — Provides direct access to Gemini text models at a 1.0x ratio with great connectivity. Choose this group when you need reliable, low-latency Gemini text generation at the standard price.

​
Important Notes

Your API key’s group controls both model access and pricing. A request will only succeed if the target model is included in your key’s assigned group. Verify your group’s supported models before deploying to production.Keep in mind that groups with lower ratios may offer discounted pricing but come with lower connectivity quality. Factor connectivity requirements into your group selection.