Anthropic-native · OpenAI-compatible

Claude and GPT behind one API
one key, one endpoint.

Call the API your SDK already knows. We handle provider connections, quota enforcement, and per-request cost accounting — while passing your stream through untouched.

Create a free account See the 60-second setup

No card needed to sign up. Keys start working once you add credit.

< 200 ms Added latency (p50)

Byte-exact SSE passthrough

Per request Token accounting

Card + bank Payment methods

Built for production traffic

Change the base URL and the key. Nothing else in your stack moves.

Anthropic-native

Messages API with tool use, prompt caching, and extended thinking — passed through, not re-serialised.

OpenAI-compatible

Chat Completions and the Responses API. Existing clients work by changing one base URL.

Streaming that actually streams

Server-sent events are forwarded chunk by chunk, never buffered. First token reaches you in milliseconds.

Cost you can audit

Every call is a row: model, input and output tokens, cache reads, amount charged. Export to CSV.

Hard spending limits

Set a cap per key. When credit runs out the request is refused with HTTP 402 — never a surprise invoice.

Provider failover

Multiple upstream channels with automatic routing and cooldown when one degrades.

Sixty seconds to your first token

Create an account, add credit, and copy one of the snippets below.

Claude Code

export ANTHROPIC_BASE_URL=https://api.apimodelhub.com
export ANTHROPIC_API_KEY=sk-spk-...

claude

curl — Anthropic Messages

curl https://api.apimodelhub.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-sonnet-5-5","max_tokens":64,
       "messages":[{"role":"user","content":"Hello"}]}'

OpenAI SDK (Python)

from openai import OpenAI

client = OpenAI(base_url="https://api.apimodelhub.com/v1",
                api_key="sk-spk-...")

Pricing

Per million tokens, billed from prepaid credit. Prices shown in VND; the equivalent in credit is listed underneath.

Model	Input / 1M	Output / 1M	Cache read / 1M
claude-haiku-4-5	468.000 ₫	2.340.000 ₫	46.800 ₫
claude-sonnet-5-5	936.000 ₫	4.680.000 ₫	93.600 ₫
claude-opus-5-5	1.872.000 ₫	9.360.000 ₫	187.200 ₫

Prompt caching is priced correctly: cache reads at ~10% of the input rate, 5-minute cache writes ×1.25, 1-hour writes ×2. The discount is passed through, not absorbed.

Credit is sold at 3.250 ₫ per $1 of credit. Add credit from your account page — pick an amount and a payment method there.

Frequently asked

Does this work with Claude Code?: Yes. Point ANTHROPIC_BASE_URL at our endpoint and use a key we issue. Tool use, prompt caching, and streaming behave exactly as with the direct API.
How do I pay?: Card via Stripe for international payments, or bank transfer with a VietQR code. Both are selected inside the checkout flow, not on this page.
Does credit expire?: No. Credit you buy does not expire and is not subject to a monthly maintenance fee.
Can I see exactly what I spent?: Yes. Your account page lists every request with model, token counts, cache tokens, and the amount charged, exportable as CSV.
What happens when I hit my limit?: The request is rejected with HTTP 402 and a message stating the remaining balance. Nothing is ever billed beyond what you prepaid.
Is this affiliated with Anthropic or OpenAI?: No. Apimodelhub is an independent service. Model names and trademarks belong to their respective owners.

Claude and GPT behind one API one key, one endpoint.