Processing or generating 100,000 words (approximately 133,300 tokens) with Llama 3.1 70B Instruct (OpenRouter) ranges from $0.0133 (cached input) to $0.0533 (full generation).
Quick answer
For 100,000 words, Llama 3.1 70B Instruct (OpenRouter) processes approximately 133,300 tokens. That costs $0.0533 as uncached input, $0.0133 with the modeled cache discount, or $0.0533 when generated as output.
Method & trust
The estimate converts the selected English word count with the model tokenizer family, then applies the published input/output price per million tokens. Real prompts vary with code, formatting, language, and system instructions.
You send 100,000 words as prompt context, documentation, or background knowledge.
Llama 3.1 70B Instruct (OpenRouter) drafts a complete 100,000 words article, chapter, or code module.
80,000 words input prompt + 20,000 words output reply.
| Model | Provider | Input Cost (100,000 words) | Cached Input Cost | Output Cost (100,000 words) | Context Limit |
|---|---|---|---|---|---|
| Llama 3.1 70B Instruct (OpenRouter) (Current) | meta | $0.0533 | $0.0133 | $0.0533 | 131,072 |
| GPT-5.6 Terra | openai | $0.256 | $0.0256 | $1.536 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $0.32 | $0.032 | $1.92 | 256,000 |
| o3-mini | openai | $0.1408 | $0.0704 | $0.5632 | 200,000 |
| o4-mini | openai | $0.1408 | $0.0352 | $0.5632 | 256,000 |
| o1-mini | openai | $0.1408 | $0.0704 | $0.5632 | 128,000 |
For Llama 3.1 70B Instruct (OpenRouter) (llama_bpe tokenizer), 100,000 words is approximately 133,300 tokens (an average ratio of 1.33 tokens per word in English). Code, technical vocabulary, and non-English scripts will have higher token densities.
Sending 100,000 words as input costs $0.0533 (or $0.0133 with prompt caching). Generating 100,000 words as output costs $0.0533. Output tokens are more expensive because autoregressive token generation requires significantly more computation per token.
Yes! If your input text is part of a repeated context, prompt caching saves 50% to 90% (costing $0.0133). For non-urgent asynchronous batch processing, 24-hour batch queues provide a flat 50% discount across all tokens (costing $0.0267 for input, $0.0267 for output).
At an average human reading speed of 250 words per minute, 100,000 words takes approximately 400 minutes to read. In contrast, Llama 3.1 70B Instruct (OpenRouter) can process or generate this text in seconds.