11-Model API Cost Comparison (4,000 Input + 1,000 Output Tokens)
| Model | Cost / 1K Calls | Monthly Bill (75,000 Calls) |
|---|---|---|
| Anthropic Claude 4 Sonnet / 3.7 Sonnet | $18.90 | $1,417.50 / mo |
| OpenAI GPT-5 / GPT-4.1 Flagship | $11.50 | $862.50 / mo |
| OpenAI GPT-4o Multimodal | $16.25 | $1,218.75 / mo |
| OpenAI GPT-4o mini (High-Speed) | $0.98 | $73.13 / mo |
| OpenAI o3-mini / o4-mini (Reasoning) | $6.33 | $474.38 / mo |
| Anthropic Claude 4 Opus | $94.50 | $7,087.50 / mo |
| Anthropic Claude 3.5 Haiku | $5.04 | $378.00 / mo |
| Google Gemini 2.5 Pro (2M Context) | $12.18 | $913.50 / mo |
| Google Gemini 2.5 Flash (1M Context) | $0.86 | $64.69 / mo |
| DeepSeek R1 / V3 Reasoning API | $1.58 | $118.50 / mo |
| Meta Llama 4 70B / 405B (Groq / Together) | $1.98 | $148.50 / mo |
Frequently Asked Questions
How much does Anthropic Claude Prompt Caching save on system prompts and RAG documents? ▾
Anthropic’s Prompt Caching reduces cached input read costs by 90%—dropping Claude Sonnet input tokens from $3.00/1M down to $0.30/1M and Claude Haiku from $0.80/1M down to $0.08/1M.