Processing or generating 500,000 words (approximately 660,000 tokens) with Claude Haiku 4.5 ranges from $0.066 (cached input) to $3.30 (full generation).
You send 500,000 words as prompt context, documentation, or background knowledge.
Claude Haiku 4.5 drafts a complete 500,000 words article, chapter, or code module.
400,000 words input prompt + 100,000 words output reply.
| Model | Provider | Input Cost (500,000 words) | Cached Input Cost | Output Cost (500,000 words) | Context Limit |
|---|---|---|---|---|---|
| Claude Haiku 4.5 (Current) | anthropic | $0.66 | $0.066 | $3.30 | 1,000,000 |
| GPT-5.6 Luna | openai | $0.128 | $0.0128 | $0.768 | 1,050,000 |
| GPT-5.4 mini | openai | $0.48 | $0.048 | $2.88 | 256,000 |
| GPT-5.4 nano | openai | $0.128 | $0.0128 | $0.80 | 128,000 |
| GPT-4o mini | openai | $0.096 | $0.048 | $0.384 | 128,000 |
| Claude 3.5 Haiku | anthropic | $0.528 | $0.0528 | $2.64 | 200,000 |
For Claude Haiku 4.5 (claude_bpe tokenizer), 500,000 words is approximately 660,000 tokens (an average ratio of 1.32 tokens per word in English). Code, technical vocabulary, and non-English scripts will have higher token densities.
Sending 500,000 words as input costs $0.66 (or $0.066 with prompt caching). Generating 500,000 words as output costs $3.30. 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 $0.066). For non-urgent asynchronous batch processing, 24-hour batch queues provide a flat 50% discount across all tokens (costing $0.33 for input, $1.65 for output).
At an average human reading speed of 250 words per minute, 500,000 words takes approximately 2000 minutes to read. In contrast, Claude Haiku 4.5 can process or generate this text in seconds.