02 / API MANUAL / REV. 01

ONE ENDPOINT.
EVERY MODEL.

OpenAI-compatible by design. Change a route name, not your client architecture.

QUICKSTART

BUY
Add a prepaid token block.
ISSUE
Create a scoped API key.
ROUTE
Send a standard chat completion.

bash

curl https://api.getargon.fun/v1/chat/completions \
  -H "Authorization: Bearer tlm_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "genai/gpt-6-astra",
    "messages": [{"role":"user","content":"Review this diff."}]
  }'

AVAILABLE ROUTES

genai/gpt-6-astra×4

GPT-6 Astra through GenAI

genai/auto×1

Auto through GenAI

genai/gemini-3.1-pro×6

Gemini 3.1 Pro through GenAI

genai/claude-fable-5×10

Claude Fable 5 through GenAI

genai/gemini-3.1-pro-preview×6

Gemini 3.1 Pro Preview through GenAI

genai/gpt-5.5×1

GPT-5.5 through GenAI

genai/claude-opus-5×12

Claude Opus 5 through GenAI

genai/claude-sonnet-5×8

Claude Sonnet 5 through GenAI

genai/gpt-5.6-luna×2

GPT-5.6 Luna through GenAI

genai/gemini-3-pro×6

Gemini 3 Pro through GenAI

genai/claude-opus-4-8×12

Claude Opus 4.8 through GenAI

genai/gpt-5.6-sol×2

GPT-5.6 Sol through GenAI

genai/claude-opus-4-8-thinking×12

Claude Opus 4.8 Thinking through GenAI

genai/gpt-5.6-terra×2

GPT-5.6 Terra through GenAI

genai/claude-opus-4-7×12

Claude Opus 4.7 through GenAI

genai/claude-sonnet-4-6×8

Claude Sonnet 4.6 through GenAI

genai/claude-haiku-4-5×4

Claude Haiku 4.5 through GenAI

NORMALIZED USAGE

Every response returns the underlying token count and its route coefficient. The ledger stays readable across providers.

json

{
  "choices": [{"message": {"role": "assistant", "content": "..."}}],
  "usage": {"input_tokens": 1840, "output_tokens": 620,
    "route_multiplier": 3, "charged_tokens": 7380}
}