What 100,000 words costs
100,000 words is about 133,000 tokens. Priced at list rates across 28 models, that is anything from $0.0133 to $1.33 for input — the same text, a 100× difference.
Every model at 100,000 words
| # | Model | Provider | Input | Output | Blended |
|---|---|---|---|---|---|
| 1 | GPT-6 Luna | OpenAI | $0.0133 | $0.0665 | $0.0299 |
| 2 | GPT-5.6 Luna | OpenAI | $0.0266 | $0.16 | $0.0665 |
| 3 | GPT-5.4 nano | OpenAI | $0.0266 | $0.166 | $0.0682 |
| 4 | Gemini 3.5 Flash-Lite | Google Gemini | $0.0399 | $0.333 | $0.123 |
| 5 | DeepSeek V4.1 Flash | DeepSeek | $0.0399 | $0.16 | $0.0798 |
| 6 | GPT-5.4 mini | OpenAI | $0.0998 | $0.599 | $0.249 |
| 7 | Gemini 3.8 Flash | Google Gemini | $0.0998 | $0.499 | $0.224 |
| 8 | Gemini 3.7 Flash | Google Gemini | $0.0998 | $0.499 | $0.224 |
| 9 | Claude Haiku 4.5 | Anthropic | $0.133 | $0.665 | $0.299 |
| 10 | Gemini 2.5 Pro | Google Gemini | $0.166 | $1.33 | $0.499 |
| 11 | Grok 4.3 | xAI Grok | $0.166 | $0.333 | $0.249 |
| 12 | DeepSeek V4 Pro | DeepSeek | $0.176 | $0.527 | $0.307 |
| 13 | Gemini 3.5 Flash | Google Gemini | $0.20 | $1.20 | $0.499 |
| 14 | GPT-5.2 | OpenAI | $0.233 | $1.86 | $0.698 |
| 15 | GPT-6.1 Sol | OpenAI | $0.266 | $1.33 | $0.599 |
| 16 | GPT-6 Sol | OpenAI | $0.266 | $1.33 | $0.599 |
| 17 | GPT-5.6 Terra | OpenAI | $0.266 | $1.60 | $0.665 |
| 18 | GPT-4.1 | OpenAI | $0.266 | $1.06 | $0.532 |
| 19 | Claude Sonnet 5 | Anthropic | $0.266 | $1.33 | $0.599 |
| 20 | Gemini 3.1 Pro (Preview) | Google Gemini | $0.266 | $1.60 | $0.665 |
| 21 | Grok 4.6 | xAI Grok | $0.266 | $0.798 | $0.466 |
| 22 | GPT-5.4 | OpenAI | $0.333 | $2 | $0.831 |
| 23 | Claude Sonnet 4.6 | Anthropic | $0.399 | $2 | $0.898 |
| 24 | GPT-5.6 Sol | OpenAI | $0.532 | $2.66 | $1.20 |
| 25 | GPT-5.5 | OpenAI | $0.665 | $3.99 | $1.66 |
| 26 | Claude Opus 5 | Anthropic | $0.665 | $3.33 | $1.50 |
| 27 | GPT-6 Astra | OpenAI | $1.33 | $6.65 | $2.99 |
| 28 | Claude Fable 5.1 | Anthropic | $1.33 | $6.65 | $2.99 |
Input = 133,000 prompt tokens. Output = 133,000 generated tokens. Blended = input + 25% of output, the same weighting used across this site.
Input vs output
Reading 100,000 words and writing 100,000 words are different line items. Output is billed at 5.0× the input rate on GPT-6 Luna and 5.0× on Claude Fable 5.1. A summarisation job pays for 100,000 words in and a fraction of that out; a generation job pays for the full volume twice.
At scale
| Documents of 100,000 words | Total tokens | Cheapest input | Priciest input |
|---|---|---|---|
| 100 | 13,300,000 | $1.33 | $133 |
| 1,000 | 133,000,000 | $13.30 | $1,330 |
| 10,000 | 1,330,000,000 | $133 | $13,300 |
| 100,000 | 13,300,000,000 | $1,330 | $133,000 |
Prompt caching changes the picture for repeated documents: cache reads bill at $0.006 per 1M on the cheapest cached tier in this dataset, a fraction of the standard input rate.
FAQ
How much does it cost to process 100,000 words with an LLM?
100,000 words is about 133,000 tokens. Input alone costs between $0.0133 on GPT-6 Luna and $1.33 on Claude Fable 5.1. Generating the same volume as output costs $0.0665 to $6.65.
What is the cheapest model for 100,000 words?
GPT-6 Luna at $0.0133 for 133,000 input tokens. It is 100.0× cheaper than Claude Fable 5.1 for the same volume.
Does that include the prompt overhead?
No. A real request also carries the system prompt and conversation history. A typical system prompt adds 200 to 800 tokens per call, which matters far more than the document size when you make thousands of calls.
Keep going
- Words to tokens converter
- Tokens to dollars
- All models ranked by price
- Recent price changes
- All 28 models ranked by price
- Run your own volumes in the calculator