DDH AI Cost Calculator

What it answers

Every reply is computed from live 2026 vendor pricing — no rough estimates, no “it depends”. The GPT picks the right model, runs the math through our public API, and quotes the source.

• “What does 1M input + 500K output tokens on GPT-5 cost me?”
• “Compare the cost of Claude Opus 4.7 vs GPT-5 for a 10K-in / 2K-out call.”
• “I run 50K calls/day on Gemini 2.5 Flash with 60% cache hits — what's my monthly bill?”
• “Cheapest frontier model for a 100K-token RAG context window?”

How it works

1. You install the GPT in one click from the ChatGPT GPT Store.
2. You ask it any pricing question — natural language, no token counting required.
3. The GPT calls our public endpoint at /api/gpt/cost which returns a structured cost breakdown sourced from the vendor's own pricing page.
4. You get the number plus a link to the matching /calc or /model page for deeper context.

{
  "model": { "id": "gpt-5", "name": "GPT-5", "brand": "OpenAI",
             "source": "https://openai.com/api/pricing" },
  "input":  { "tokens": 1000, "pricePer1m": 1.25, "cost": 0.001250 },
  "output": { "tokens": 500,  "pricePer1m": 10,   "cost": 0.005000 },
  "batch_applied": false,
  "total_cost_usd": 0.00625,
  "projections": {
    "per_1k_calls_usd":   6.25,
    "per_100k_calls_usd": 625,
    "per_1m_calls_usd":   6250
  },
  "summary": "GPT-5: 1,000 input + 500 output tokens = $0.006250 per call. Source: https://openai.com/api/pricing.",
  "permalink": "https://aipromptshub.co/model/gpt-5",
  "attribution": "Free to use. Credit \"AI Prompts Hub\" + link to permalink. CC-BY-4.0."
}

Also available as

For builders — call the API directly

The same endpoint powers all three integrations. CORS-open, no auth, edge-cached one hour.

curl -X POST https://aipromptshub.co/api/gpt/cost \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5","input_tokens":1000,"output_tokens":500}'

OpenAPI 3.1 spec: /api/gpt/openapi.yaml. Full model list: /api/gpt/cost/models.