NVIDIA Nemotron Nano 12B v2 VL BF16: pay per request with x402
Call NVIDIA Nemotron Nano 12B v2 VL BF16 with no account or API key. Pay per request in USDC on Base via x402.
Pay per request
Send a request to ask.singitai.app, get HTTP 402 with the exact price, pay in USDC on Base, get the answer. No signup, no monthly plan.
Provider: NVIDIA · Context window: 128,000 tokens · Tags: vision
Price
| USD per 1M tokens | Input | Output |
|---|---|---|
| SingIt price (expected, markup included) | $0.00286 | $0.00858 |
| List price of the model (for reference) | $0.2 | $0.6 |
The expected price is estimate from sellers' quotes, not yet confirmed by answers we paid for, so the real price can be higher. It is an expected price, not a guarantee; you pay what each answer actually used.
Price = what the provider charges for the answer + 30% markup + a settlement fee per request ($0.001 on Base, $0.002 on Solana).
Per request, reference request
One short message, max_tokens 800 (30 characters), 12 input tokens. upto settles actual usage, so a real answer normally costs less; exact is flat, with no refund.
| Network | upto ceiling (USD) | exact flat price (USD) |
|---|---|---|
| Base | $0.0015 | $0.0015 |
| Solana | $0.00202 | $0.002007 |
Prices as of 2026-10-30 19:50:05 UTC (snapshot of ask.singitai.app/v1/models and /pricing.json). Live: /v1/models and /pricing.json. Live values are authoritative and the 402 response states the exact price.
Call it with x402
curl -i https://ask.singitai.app/v1/chat/completions \
-H 'Content-Type: application/json' \
-d '{"model":"nvidia-nemotron-nano-12b-v2","max_tokens":200,"messages":[{"role":"user","content":"Hello"}]}'The unpaid POST returns HTTP 402 and a PAYMENT-REQUIRED header: base64 JSON, x402 v2. Sign one offer (upto first, then exact; Base then Solana) and retry the same request with the PAYMENT-SIGNATURE header. HTTP 200 returns choices[0].message.content and billing.totalAtomic.
Follow the first paid call walkthrough, developer guide and llms.txt.
For a free try, choose from 6 free models, for example free/deepseek-v4.1-flash: 20 answers a day per network, no wallet. The shared daily cost cap is $0.25; free answers may run out for the day (free_budget_exhausted) and return at midnight UTC.
Use it from an agent
MCP: https://app.singitai.app/mcp. Free tools without a key: list_models, get_price, list_price_symbols.
The elizaOS plugin @singitai/plugin-singit has a paid-chat action that pays on Base. Its per-call and per-day caps are held in the plugin's memory: in-process, reset on restart, not on-chain, not enforced by SingIt's servers. It does not use AgentAllowance.
Give your agent a budget
Agent spending limits for USDC on Base · How to limit what an AI agent can spend · x402 payments with a budget for AI agents
FAQ
What is x402?
- x402 is an HTTP payment protocol. An unpaid request returns HTTP 402 with a payment offer; sign one offer and retry the same request with proof of payment.
Do I need an account?
- No account or API key is needed for these x402 calls.
Which network? Base?
- Base or Solana, paid in USDC. This page focuses on Base.