What one run costs
A typical run of this task sends about 12K tokens and gets back 3.5K, over 3 calls — 46.5K tokens in total. Those are the numbers below, against the published prices.
| Model | Vendor | Per run | Can do what this needs? | Price evidence |
|---|---|---|---|---|
| GPT-5 Nano | OpenAI | $0.00600 | not stated | Vendor page |
| Gemini 2.5 Flash Lite | $0.00780 | not stated | Vendor page | |
| GPT-4.1 nano | OpenAI | $0.00780 | confirmed | Corroborated |
| Claude Haiku 5.5 | Anthropic | $0.00885 | confirmed | Vendor page |
| GPT-6 Luna | OpenAI | $0.00885 | not stated | Vendor page |
| Jamba Mini | AI21 | $0.0114 | confirmed | Vendor page |
| GPT-4o mini | OpenAI | $0.0117 | confirmed | Corroborated |
| GPT-5.6 Luna | OpenAI | $0.0198 | not stated | Vendor page |
Ask for the outline first. A model given a document and asked for slides will mirror the document’s structure, which is almost never the structure of a good talk. Argue with the outline, then generate from the agreed one.
Three calls rather than one: outline, slides, then a pass for the speaker notes. Each is cheap, and the iteration is what produces something worth presenting.
No model on this site makes the slide file itself. They produce the content; something else assembles it.
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.