What it's for
The API behind our micro-SaaS, agent and RAG notes.
- AI features inside your own app
- Agents and RAG
- Embeddings and speech
Plans and prices
| Plan | Price (USD unless shown) | What you get |
|---|---|---|
| gpt-6.1-sol | $2.00 input / $10.00 output per 1M tokens | Short context, standard tier |
| gpt-6-luna | $0.10 input / $0.50 output per 1M tokens | Short context, standard tier |
| gpt-6-astra | $10.00 input / $50.00 output per 1M tokens | Short context, standard tier |
Read on the official pricing page on . Prices change; check before you buy.
Free plan in real output
No free plan; pay per token.
- A support-bot answer of about 1,500 input and 400 output tokens on gpt-6-luna costs 1,500 × $0.10/1M + 400 × $0.50/1M = $0.00035, so about 2,850 answers per $1.
- The same answer on gpt-6.1-sol: 1,500 × $2/1M + 400 × $10/1M = $0.007, so about 140 answers per $1.
Cost at your stage
| Plan | Monthly cost | |
|---|---|---|
| Month 1 (testing, low volume) | Pay as you go | $5 |
| Month 12 (working volume) | Pay as you go | $50 |
Break-even: Charge per seat or per month, not per answer. A $29/mo micro-SaaS customer covers about 4,100 gpt-6.1-sol answers at the cost above.
Month-12 plan is our estimate of what a solo seller at steady weekly output needs, based on the plan limits above. Costs exclude tax.
Commercial use and output rights
You own or may use output commercially
“As between Customer and OpenAI, to the extent permitted by applicable law, Customer: (a) retains all ownership rights in Input; and (b) owns all Output. OpenAI hereby assigns to Customer all OpenAI's right, title, and interest, if any, in and to Output.”
Your data and AI training: 'OpenAI will only use Customer Content as necessary to provide Customer with the Services, comply with applicable law, enforce the OpenAI Policies, and prevent abuse. OpenAI will not use Customer Content to develop or improve the Services, unless Customer explicitly agrees to such use.' (Services Agreement, 4.2)
This is a summary of public terms, not legal advice. Terms change; re-read the clause before client work.
First-hour setup
- Create a platform account and set a monthly spend limit.
- Create a project-scoped API key and store it in an environment variable.
- Start with the cheapest model and measure quality.
- Log tokens per request.
- Price your product off the measured cost per user.
Competitors compared
| Tool | Pricing model | Free plan | Starting price | Difficulty | Output rights | Best for |
|---|---|---|---|---|---|---|
| OpenAI API | Pay as you go | No | $2.00 input / $10.00 output per 1M tokens | 3/5 | You own or may use output commercially | Building AI features into your own app |
| Claude API | Pay as you go | No | See official site | 3/5 | – | Long-context and coding features in apps |
| Gemini API | Pay as you go | Yes | See official site | 3/5 | – | Cheap experiments with a free tier |
| OpenRouter | Pay as you go | Yes | See official site | 3/5 | – | Switching between models without new accounts |
| Groq | Pay as you go | Yes | See official site | – | – | Low-latency open-model inference |
| Ollama | Free, open source | Yes | Free | 2/5 | – | Private, offline models |
| Together AI | Pay as you go | No | See official site | – | – | Running and fine-tuning open models |
Learning difficulty
3/5: Needs programming and cost monitoring.
When not to use it
Not needed for one-off writing work; ChatGPT is cheaper and simpler for that.
Cheaper or free alternatives:
- Claude API: Long-context and coding features in apps (no free plan)
- Gemini API: Cheap experiments with a free tier (yes free plan)
- OpenRouter: Switching between models without new accounts (yes free plan)
- Groq: Low-latency open-model inference (yes free plan)
Which of our hustles use it
- AI customer-service chatbots for small shops: The language model behind custom bots
- Micro-SaaS on an LLM API: The model API
- AI agent building for clients: Models and the Agents SDK for tool-calling agents
- RAG knowledge-base chatbots for companies: Embeddings, file search and the answer model
Price history
Each monthly check adds rows here. Older rows are never edited, so you can see when prices changed.
Sources
- OpenAI API pricing (read 17 October 2026)
- OpenAI Services Agreement (read 17 October 2026)
Last verified . Anything we couldn't confirm on an official page is marked "not published", "not shown" or "not verified".