Ask Janek AI with Janek

All questions

Make.com Zero to Hero 2026 – 2.4 AI Basics in Make · published 9 November 2026

Play at 7:44 Watch on YouTube AI

What do the OpenAI and xAI prices on screen actually mean?

The caption prompt was 19 tokens and 90 characters. On the OpenAI price page, GPT-6 Astra is $10 in and $50 out per million tokens, GPT-6.1 Sol is $2 in and $10 out, and GPT-6 Luna is 10 cents in and 50 cents out. Grok 4.7 is $2 in and $6 out per million, with a context of 500,000 tokens. Those text prices are still on the pages. They change, so I check the page again before I decide.

The tokenizer is at platform.openai.com/tokenizer. It needs no login. Each color is one token. Short words such as a, for, and about are one token. Fit Fuel is split into fit and fuel because it is not an ordinary word. Token ID shows the numbers the model uses. Input is the prompt. Output is the answer, and it usually costs more per token. Cached input is cheaper when the same text is sent again. Batch is 50% cheaper if I do not need the answer immediately. Nineteen tokens plus a short answer is a tiny fraction of a cent even on the bigger models. On the xAI models page, images start at 2 cents and video starts at 2 cents a second. That 2-cent video rate is the cheapest start. grok-imagine-video itself is still 5 cents a second at 480p and 7 cents at 720p, which is How do I generate a Grok video in Make.com, and what does it cost? The same Grok 4.7 page charges more once a prompt passes 200,000 tokens. Context is how much text the model can look at in one go. Further down, the page says which model is for code, chat, images, video, and voice, and the aliases are other names for the same models.