Your TokenPlan works on every model on AgentMarket with one FluxA API key. The model list below shows the price per 1M tokens for each.
AgentMarket TokenPlan
Choose a balance and get a FluxA API key. Use it on every model on AgentMarket, including Claude, GPT, Gemini and the GLM and DeepSeek models from Baidu AI Cloud, at the prices in the model list.
Standard
Advanced
- Per-token prices for every model are in the Model list tab below.
- Works with Claude Code, OpenCode, and other mainstream developer tools.
Claude Opus 5.5
Claude Sonnet 5.5
Claude Fable 5.1
GPT 6.1 Sol
GPT 6 Astra
Gemini 4 Argon
Grok 4.7
GLM 5.3
DeepSeek V4 Pro
After you buy, create a FluxA API key on the API keys page. Send requests to either endpoint below with that key, and pick any model from the Model list.
Endpoints
The same key works on both. Use the Anthropic one for Claude Code and other Anthropic-protocol tools, and the OpenAI one for everything else. A tool that asks for a base URL wants the part before /v1/messages or /chat/completions.
Claude Code
Claude Code adds /v1/messages itself, so it takes the Anthropic endpoint without that part.
export ANTHROPIC_BASE_URL="https://router.fluxapay.xyz/llm" export ANTHROPIC_AUTH_TOKEN="<your-fluxa-api-key>" export ANTHROPIC_MODEL="anthropic/claude-sonnet-5-5"
Try it
Anthropic protocol, with a Claude model:
curl https://router.fluxapay.xyz/llm/v1/messages \
-H "x-api-key: <your-fluxa-api-key>" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{"model":"anthropic/claude-sonnet-5-5","max_tokens":1024,"messages":[{"role":"user","content":"Hello"}]}'OpenAI protocol, with a GPT model:
curl https://router.fluxapay.xyz/llm/v1/chat/completions \
-H "Authorization: Bearer <your-fluxa-api-key>" \
-H "Content-Type: application/json" \
-d '{"model":"openai/gpt-6.1-sol","messages":[{"role":"user","content":"Hello"}]}'