What one run costs
A typical run of this task sends about 2K tokens and gets back 600, over 3 calls — 7.8K tokens in total. Those are the numbers below, against the published prices.
| Model | Vendor | Per run | Price evidence |
|---|---|---|---|
| GPT-5 Nano | OpenAI | $0.00102 | Vendor page |
| Gemini 2.5 Flash Lite | $0.00132 | Vendor page | |
| GPT-4.1 nano | OpenAI | $0.00132 | Corroborated |
| Claude Haiku 5.5 | Anthropic | $0.00150 | Vendor page |
| GPT-6 Luna | OpenAI | $0.00150 | Vendor page |
| Jamba Mini | AI21 | $0.00192 | Vendor page |
| GPT-4o mini | OpenAI | $0.00198 | Corroborated |
| GPT-5.6 Luna | OpenAI | $0.00336 | Vendor page |
The reason these are recognisable is not the model. It is that people ask for “a LinkedIn post about X” and get the average of every LinkedIn post ever written.
Paste three things you have actually written and ask it to match the voice. Then delete the opening line it produces, which is reliably a throat-clear, and the closing question, which is reliably a plea for engagement.
Three cheap passes beat one expensive one: draft, cut by a third, then fix the rhythm.
What we would pick
This section is our judgement, not a figure read off a page. Everything above is arithmetic on published prices; this is an opinion, and it is labelled as one.