Skip to main content

​
Request

POST https://ninjachat.ai/api/v1/estimate
Content-Type: application/json
No API key required for basic estimates. Pass your key if you want balance-aware responses.
{
  "model": "claude-sonnet-4.6",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Write a 500-word essay about climate change."}
  ],
  "count": 1000
}

​
Response

{
  "model": "claude-sonnet-4.6",
  "model_name": "Claude Sonnet 4.6",
  "provider": "Anthropic",
  "tier": "premium",
  "cost_per_request_cents": 1.5,
  "cost_per_request": "$0.0150",
  "estimated_tokens": {
    "prompt": 28,
    "completion": 11
  },
  "for_count": 1000,
  "total_cost_cents": 1500.0,
  "total_cost": "$15.0000",
  "cheaper_alternatives": [
    {
      "id": "claude-sonnet-4.5",
      "name": "Claude Sonnet 4.5",
      "cost_cents": 0.6,
      "savings_percent": 60,
      "capabilities": ["text", "code", "analysis", "reasoning"]
    },
    {
      "id": "gpt-5",
      "name": "GPT-5",
      "cost_cents": 0.6,
      "savings_percent": 60,
      "capabilities": ["text", "code", "reasoning", "analysis", "math"]
    },
    {
      "id": "deepseek-v3",
      "name": "DeepSeek V3",
      "cost_cents": 0.3,
      "savings_percent": 80,
      "capabilities": ["text", "code", "math", "reasoning"]
    }
  ],
  "monthly_estimate": {
    "at_100": "$0.15",
    "at_1000": "$1.50",
    "at_10000": "$15.00",
    "at_100000": "$150.00"
  },
  "ensemble_pricing": {
    "ensemble": { "cost_cents": 4, "cost": "$0.040" },
    "ensemble-quality": { "cost_cents": 5, "cost": "$0.050" }
  },
  "note": "Token estimates are approximate. Actual billing is per-request, not per-token."
}

​
Parameters

ParameterTypeRequiredDefaultDescription
modelstringYes—Any model ID (concrete models only, not auto*).
messagesarrayNo—Used for approximate token estimation.
countintegerNo1Number of requests to estimate total cost for (1–1,000,000).
modelsarrayNo—Compare pricing across multiple models simultaneously.

​
Compare multiple models at once

Pass models to get a side-by-side pricing comparison: 
{
  "model": "claude-sonnet-4.6",
  "count": 50000,
  "models": ["gpt-5", "claude-sonnet-4.6", "deepseek-v3", "gpt-5-mini", "claude-opus-4.6"]
}

{
  "model_comparison": [
    { "model": "gpt-5-mini",      "cost_cents": 0.3, "total_cost_cents": 15000 },
    { "model": "deepseek-v3",     "cost_cents": 0.3, "total_cost_cents": 15000 },
    { "model": "gpt-5",           "cost_cents": 0.6, "total_cost_cents": 30000 },
    { "model": "claude-sonnet-4.6","cost_cents": 1.5, "total_cost_cents": 75000 },
    { "model": "claude-opus-4.6", "cost_cents": 3.0, "total_cost_cents": 150000 }
  ]
}

​
Code examples

import requests

def estimate_batch_cost(model: str, count: int) -> float:
    """Returns estimated cost in cents. Free to call."""
    r = requests.post("https://ninjachat.ai/api/v1/estimate",
        json={"model": model, "count": count}
    )
    return r.json()["total_cost_cents"]

# Check cost before running a large batch
cost = estimate_batch_cost("claude-sonnet-4.6", 10000)
budget = 5000  # 50 dollars in cents

if cost > budget:
    print(f"Too expensive: ${cost/100:.2f}. Try 'gpt-5' or 'deepseek-v3'.")
else:
    print(f"Within budget: ${cost/100:.2f}. Proceeding with batch.")
    # ... run batch

​
Monthly cost planning

The monthly_estimate field in every response gives you cost projections at 4 scale levels: 
"monthly_estimate": {
  "at_100":    "$0.06",
  "at_1000":   "$0.60",
  "at_10000":  "$6.00",
  "at_100000": "$60.00"
}
Use this to plan credit purchases and set spend alerts in your application.