chatgpt · openai · chat & research
ChatGPT, minus the second subscription.
OpenAI's GPT-6 family for reasoning, everyday writing and images, in the workspace where the rest of your models already live.
- made by
- OpenAI
- versions
- 4
- from
- 1 token
- included on
- Starter+
best at
What ChatGPT is good at
ChatGPT is the generalist most people start with, and for good reason. It is quick to understand loosely worded requests, comfortable switching from a cover letter to a spreadsheet formula in the same thread, and decisive in a way that suits first drafts, brainstorming and the hundred small questions a working day throws up.
The GPT-6 models are at their best on structured reasoning: breaking a messy problem into steps, checking arithmetic, planning a project, or turning a vague brief into a list of concrete tasks. GPT-6 Astra is the one to reach for when a problem has several moving parts and you want them all accounted for.
In LazySusan the ChatGPT tool also does more than text. Switch on Image and the same conversation produces a picture with GPT Image 2.5, switch on Search when the answer depends on current information, or turn on Research for a slower, cited report built from many sources. Everything lands in the same history as your other chats, so the image, the draft and the follow-up questions stay together.
versions
ChatGPT versions on LazySusan
| Version | Use it for |
|---|---|
| GPT-6.1 Sol | The default. Smart and fast, and the right starting point for almost everything. |
| GPT-6 Astra | OpenAI's most powerful model, for hard reasoning, planning and analysis where accuracy matters more than speed. |
| GPT-6 Luna | The fastest and cheapest GPT-6, for short answers, classification and quick rewrites. |
| GPT-5.5 | The previous flagship, kept for anyone who prefers its style or needs consistency with older work. |
pricing
What ChatGPT costs on LazySusan
| Generation | Tokens |
|---|---|
| GPT-6 Luna1–2 from a short to a long reply | 1 |
| GPT-6.1 Sol7–40 from a short to a long reply | 16 |
| GPT-5.519–117 from a short to a long reply | 45 |
| GPT-6 Astra35–197 from a short to a long reply | 77 |
| GPT Image 2.5, one square image at medium quality | 64 |
Chat costs are for a message of about 300 characters; you pay for what the reply actually uses. A typical GPT-6.1 Sol reply is 16 tokens, so Starter's 1,000 monthly tokens go a long way on everyday questions. Prompts longer than 272,000 tokens cost more per token on GPT-6 models.
Every plan includes ChatGPT. Starter is $8 a month with 1,000 tokens; unused tokens roll over while you're subscribed. Compare plans.
prompts
Prompts to try with ChatGPT
// planning
I'm moving my small bakery's orders from paper to a spreadsheet. Plan the columns, the formulas I'll need for daily totals, and a checklist for the first week.
// email
Write a polite but firm reply to a client who has missed two invoice deadlines. Offer a payment plan and keep it under 120 words.
// reasoning
We have 3 developers and 11 tickets with these estimates and dependencies. Propose a two-week schedule and tell me which ticket is the real bottleneck.
// learning
Explain how a mortgage offset account works with a worked example, then quiz me with three questions to check I understood.
// image
With Image on: a flat-lay photo of a home espresso setup on a walnut counter, morning light, a handwritten recipe card in the corner.
tips
Getting better results
- 01Stay on GPT-6.1 Sol by default and move to GPT-6 Astra only for the problems that need it. A typical Astra reply is 77 tokens, against 16 tokens for Sol.
- 02GPT-6 Luna is ideal for bulk, low-stakes jobs such as tagging, short summaries or tidying text. At around 1 token a reply you can run it freely.
- 03Turn on Search when a question depends on prices, schedules or news. Without it, the model answers from what it learned in training.
- 04Image quality changes the cost: a square medium-quality image is 64 tokens, high quality 254 tokens. Medium is usually enough to judge a concept.
- 05Smart prompt rewrites a rough request into a clearer one before it's sent. It adds 5 tokens and is most useful when you're not sure how to ask.
alternatives
When to pick a different tool
- Claude if you're working through a long document or want a careful edit that keeps your voice.
- Perplexity if you need every claim backed by a source you can open.
- GPT Image if you want full control over image quality, shape and edits rather than a picture in the middle of a chat.
- DeepSeek if you have lots of routine reasoning to get through and cost matters most.
also on the api
ChatGPT in your own app
Call GPT-6 with the same key and request shape you use for Claude, Gemini or an image model.
One key, our SDKs for JavaScript and Python, prepaid credits with no subscription. LazySusan API · docs
faq
ChatGPT on LazySusan: questions
Yes. Every LazySusan plan includes OpenAI's current models, starting with Starter at $8 a month. You don't need an OpenAI account, and the same plan covers Claude, Gemini, Grok and the image, video and voice tools.
GPT-6.1 Sol, GPT-6 Astra, GPT-6 Luna and GPT-5.5. You choose the model in the conversation and can change it between messages.
It uses tokens from your plan. A typical message costs about 1 token on GPT-6 Luna, 16 tokens on GPT-6.1 Sol, 45 tokens on GPT-5.5 and 77 tokens on GPT-6 Astra. Long replies, attachments and web search add to that.
Yes. Switch on Image in the ChatGPT composer and it creates a picture with GPT Image 2.5. A square image at medium quality is 64 tokens. For more control, the dedicated GPT Image tool adds editing and more shapes.
Yes. Attach PDFs, Word, Excel and PowerPoint files, text files or images and ask about them. Reading the file counts toward that message's cost.
Yes. Import your ChatGPT conversations into LazySusan and they're searchable next to your new chats.
Yes. GPT-6.1 Sol, GPT-6 Astra and GPT-6 Luna are in the LazySusan developer API, with one key and our SDK for JavaScript and Python. The API runs on its own prepaid balance with no subscription.
get started
ChatGPT and every other model, one plan.
One login, one bill and one Library for everything you make, with the right model for each part of the job.
Starter is $8/month. Cancel anytime.