Models & routing

Tokelix supports OpenAI and Anthropic compatible API formats. Point the official SDK base URL to Tokelix and connect GPT, Claude and DeepSeek through one gateway. Available models, prices and protocols depend on configured channels.

Configuration and access

Each API key is bound to an organization and member. Wallet balance and organization and member budgets govern new requests. Input, output, cache read and cache write use the request price snapshot.

Error handling

401 — Invalid or expired key
403 — Membership, balance or budget restriction
429 — Rate limit; retry with backoff
503 — No route matches the model, protocol or price cap

Model	Context	Input (List)	Output (List)	Input (Tokelix)	Output (Tokelix)	Provider
gpt-5-6-lunaOpenAI	—	—per 1M tokens	—	$0.042 Cache $0.042	$0.336	OpenAI
deepseek-v4-1-flashDeepSeek	—	$0.3per 1M tokens	$1.2	$0.12 Cache $0.12	$0.474996	DeepSeek
gemini-3-8-flash-openaiGoogle	—	—per 1M tokens	—	$0.225 Cache $0.225	$1.125	Google
gemini-3-7-flash-openaiGoogle	—	—per 1M tokens	—	$0.225 Cache $0.225	$1.125	Google
gemini-3-6-flash-openaiGoogle	—	—per 1M tokens	—	$0.225 Cache $0.225	$1.125	Google
claude-haiku-4-5Anthropic	—	$1per 1M tokens	$5	$0.275 Cache $0.275	$1.375	Anthropic
gemini-3-5-flash-openaiGoogle	—	—per 1M tokens	—	$0.45 Cache $0.45	$2.7	Google
gpt-5-6-terraOpenAI	—	—per 1M tokens	—	$0.42 Cache $0.42	$3.36	OpenAI