Cheapest GPT-5.6 Luna API: Tokemart is 94% below the official price
GPT-5.6 Luna costs $0.200/$1.20 per 1M tokens from OpenAI. The cheapest quote is Tokemart: $0.011 in, $0.069 out. Full table and buying checklist below.
| # | Provider | 1M input | 1M output | Discount | Tokens per $100 |
|---|---|---|---|---|---|
| — | Official list | $0.200 | $1.20 | 0% | 273M |
| 1 | Tokemartdemo | $0.011 | $0.069 | −94% | 4.78B |
| 2 | SwiftRoutedemo | $0.018 | $0.105 | −91% | 3.11B |
| 3 | PromptPortdemo | $0.022 | $0.133 | −89% | 2.46B |
| 4 | ModelMilldemo | $0.029 | $0.176 | −85% | 1.86B |
| 5 | APIBridgedemo | $0.062 | $0.369 | −69% | 887M |
| 6 | FluxAPIdemo | $0.096 | $0.573 | −52% | 571M |
| 7 | RelayOnedemo | $0.124 | $0.745 | −38% | 440M |
| 8 | ByteGatedemo | $0.138 | $0.826 | −31% | 396M |
Tokens per $100 assume a “code” mix: 5 input tokens per output token. Prices are crawled; verify them on the provider page before buying.
What to check before buying
- Test the key with the free API checker — it verifies the endpoint responds, without sending tokens.
- Compare the blended cost for your workload, not just the input price: output tokens dominate in agent workloads.
- Check uptime and minimum top-up: the cheapest provider is useless if it is down half the day.