Processing or generating 2,500 words (approximately 3,333 tokens) with Together AI — DeepSeek V4 ranges from $0.000417 (cached input) to $0.005 (full generation).
You send 2,500 words as prompt context, documentation, or background knowledge.
Together AI — DeepSeek V4 drafts a complete 2,500 words article, chapter, or code module.
2,000 words input prompt + 500 words output reply.
| Model | Provider | Input Cost (2,500 words) | Cached Input Cost | Output Cost (2,500 words) | Context Limit |
|---|---|---|---|---|---|
| Together AI — DeepSeek V4 (Current) | together | $0.001667 | $0.000417 | $0.005 | 1,000,000 |
| GPT-5.6 Terra | openai | $0.0064 | $0.00064 | $0.0384 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $0.008 | $0.0008 | $0.048 | 256,000 |
| o3-mini | openai | $0.00352 | $0.00176 | $0.0141 | 200,000 |
| o4-mini | openai | $0.00352 | $0.00088 | $0.0141 | 256,000 |
| o1-mini | openai | $0.00352 | $0.00176 | $0.0141 | 128,000 |
For Together AI — DeepSeek V4 (deepseek_bpe tokenizer), 2,500 words is approximately 3,333 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 2,500 words as input costs $0.001667 (or $0.000417 with prompt caching). Generating 2,500 words as output costs $0.005. 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 $0.000417). For non-urgent asynchronous batch processing, 24-hour batch queues provide a flat 50% discount across all tokens (costing $0.000833 for input, $0.0025 for output).
At an average human reading speed of 250 words per minute, 2,500 words takes approximately 10 minutes to read. In contrast, Together AI — DeepSeek V4 can process or generate this text in seconds.