Together AI through an OpenAI-compatible gateway: Llama, DeepSeek and Qwen cost tracking
Route Together AI's open-weight models (Llama, DeepSeek, Qwen and more) through one LLM gateway with the OpenAI SDK. Any org/model name is detected automatically, with per-feature cost and budgets.
Together AI hosts hundreds of open-weight models behind an OpenAI-compatible API. Cognocient AI forwards calls as sent, with your Together key.
Routing rule: anything with a slash
Any model name of the form org/model goes to Together AI, so new model families work the day Together lists them. Exceptions:
vertex/...goes to Vertex AI andcustom/...to your self-hosted endpoint.openai/gpt-oss-*and lower-caseqwen/qwen3*go to Groq. Together's own capitalisedQwen/Qwen3-...names still go to Together.
Open-weight models and cost per customer
Open-weight models are often used to serve a SaaS feature at a margin. Tag calls with X-Cost-Customer and Customer Margin (Growth plan) shows the AI cost of each of your customers next to what they pay you.
Related articles
Supported Providers
OpenAI, Anthropic, Gemini, Mistral, Groq, Together, Azure, AWS Bedrock, Vertex AI, and custom OpenAI-compatible endpoints.
OpenAI
OpenAI SDK through the gateway: chat, Responses, embeddings, images, audio and batches with per-feature cost and budgets.
Anthropic (Claude)
Claude through the OpenAI SDK (translated) or the native Anthropic SDK (passthrough), with prompt-cache-aware cost.