Processing or generating 5,000 words (approximately 6,665 tokens) with Microsoft Phi-4 (14B) ranges from $0.000167 (cached input) to $0.002 (full generation).
You send 5,000 words as prompt context, documentation, or background knowledge.
Microsoft Phi-4 (14B) drafts a complete 5,000 words article, chapter, or code module.
4,000 words input prompt + 1,000 words output reply.
| Model | Provider | Input Cost (5,000 words) | Cached Input Cost | Output Cost (5,000 words) | Context Limit |
|---|---|---|---|---|---|
| Microsoft Phi-4 (14B) (Current) | microsoft | $0.000667 | $0.000167 | $0.002 | 16,384 |
| GPT-5.6 Luna | openai | $0.00128 | $0.000128 | $0.00768 | 1,050,000 |
| GPT-5.4 mini | openai | $0.0048 | $0.00048 | $0.0288 | 256,000 |
| GPT-5.4 nano | openai | $0.00128 | $0.000128 | $0.008 | 128,000 |
| GPT-4o mini | openai | $0.00096 | $0.00048 | $0.00384 | 128,000 |
| Claude Haiku 4.5 | anthropic | $0.0066 | $0.00066 | $0.033 | 1,000,000 |
For Microsoft Phi-4 (14B) (cl100k_base tokenizer), 5,000 words is approximately 6,665 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 5,000 words as input costs $0.000667 (or $0.000167 with prompt caching). Generating 5,000 words as output costs $0.002. 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 $0.000167). For non-urgent asynchronous batch processing, 24-hour batch queues provide a flat 50% discount across all tokens (costing $0.000333 for input, $0.001 for output).
At an average human reading speed of 250 words per minute, 5,000 words takes approximately 20 minutes to read. In contrast, Microsoft Phi-4 (14B) can process or generate this text in seconds.