A standard 500 pages document (~250,000 words / 333,250 tokens) costs $0.4999 to ingest and $0.5014 to generate a comprehensive executive summary using NVIDIA Nemotron-4 340B.
Feed the entire 500 pages document into context to extract fields, entities, or answer queries.
Ingest 500 pages and generate a structured 500-token executive summary with key takeaways.
Translate or reformat the entire 500 pages document into an equivalent length output.
| Model | Provider | Ingestion (500 pages) | Cached Ingestion | Summary Cost | Context Limit |
|---|---|---|---|---|---|
| NVIDIA Nemotron-4 340B (Current) | nvidia | $0.4999 | $0.125 | $0.5014 | 128,000 |
| GPT-5.6 Sol | openai | $1.60 | $0.16 | $1.615 | 1,050,000 |
| GPT-5.6 Cyber | openai | $4.00 | $0.40 | $4.038 | 1,050,000 |
| GPT-5.5 Standard | openai | $1.60 | $0.16 | $1.615 | 512,000 |
| GPT-5.5 Pro | openai | $9.60 | $0.96 | $9.69 | 512,000 |
| o3-pro (Frontier Reasoning) | openai | $6.40 | $0.64 | $6.44 | 1,000,000 |
Assuming a standard single-spaced document with ~500 words per page, a 500 pages document contains approximately 250,000 words, which translates to roughly 333,250 tokens using NVIDIA Nemotron-4 340B's tokenizer. Dense PDF tables, legal boilerplate, and OCR scans may increase this by 20–30%.
NVIDIA Nemotron-4 340B has a context window of 128,000 tokens. A 500 pages document consumes only 260.35% of its total window, easily fitting in a single prompt without requiring chunking or vector search.
Ingesting the document and outputting a concise 500-token executive summary costs $0.5014. Using 24-hour async batch API queues drops this cost to $0.2507.
If your workflow repeatedly queries or chats with this same 500 pages document, prompt caching reduces subsequent turn input costs by up to 90% (to $0.125 per turn instead of $0.4999).