Perffeco API

Programmatic access to real-time LLM pricing, GPU cloud costs, benchmark data, and AI agent economics. Build cost-aware AI applications with live data.

Authentication

All API requests require an API key passed via the Authorization header. You can generate API keys from your account dashboard (Team plan and above).

curl -H "Authorization: Bearer YOUR_API_KEY" \
  https://api.perffeco.com/v1/llm/pricing

Base URL

All endpoints are relative to:

https://api.perffeco.com/v1

Rate Limits

Rate Limits by Plan

Team: 10,000 requests/month · Enterprise: Unlimited · Burst limit: 60 requests/minute

LLM Pricing

GET /llm/pricing

Returns current pricing for all tracked LLM models including input/output token costs, cached pricing, and batch API pricing.

Parameter	Type	Required	Description
provider	string	optional	Filter by provider: openai, anthropic, google, mistral, deepseek, xai
model	string	optional	Filter by model name (partial match)
sort	string	optional	Sort by: input_price, output_price, name. Default: input_price

Response

{
  "data": [
    {
      "model": "gpt-4o",
      "provider": "openai",
      "input_price_per_1m": 2.50,
      "output_price_per_1m": 10.00,
      "cached_input_price_per_1m": 1.25,
      "batch_input_price_per_1m": 1.25,
      "batch_output_price_per_1m": 5.00,
      "context_window": 128000,
      "max_output": 16384,
      "updated_at": "2026-03-14T00:00:00Z"
    }
  ],
  "meta": {
    "total": 22,
    "updated": "2026-03-14T06:00:00Z"
  }
}

GET /llm/cost-per-task

Returns estimated cost per task type (chat, summarisation, code generation, translation, classification, RAG) for each model.

Parameter	Type	Required	Description
task	string	optional	Filter: chat, summarise, code, translate, classify, rag

Response

{
  "data": [
    {
      "model": "deepseek-v3",
      "tasks": {
        "chat": 0.0003,
        "summarise": 0.0005,
        "code": 0.0008,
        "translate": 0.0004,
        "classify": 0.0001,
        "rag": 0.0006
      },
      "currency": "USD"
    }
  ]
}

GPU Pricing

GET /gpu/pricing

Returns GPU cloud pricing across all tracked providers and GPU models.

Parameter	Type	Required	Description
gpu	string	optional	Filter: h100, a100, h200, b200, l40s, rtx4090
provider	string	optional	Filter by provider name
type	string	optional	Filter: on_demand, spot, reserved

Response

{
  "data": [
    {
      "gpu": "H100 SXM 80GB",
      "provider": "vast.ai",
      "on_demand_hourly": 1.49,
      "spot_hourly": 1.19,
      "vram_gb": 80,
      "type": "marketplace",
      "updated_at": "2026-03-14T00:00:00Z"
    }
  ],
  "meta": { "total": 32, "providers": 12 }
}

Benchmarks

GET /benchmarks

Returns benchmark scores including CIS (Composite Intelligence Score), Arena Elo, GPQA, SWE-bench, MATH-500, and HumanEval.

Parameter	Type	Required	Description
model	string	optional	Filter by model name
benchmark	string	optional	Filter: cis, arena_elo, gpqa, swe_bench, math_500, humaneval
sort	string	optional	Sort by any benchmark field. Default: cis

Response

{
  "data": [
    {
      "model": "claude-opus-4-6",
      "provider": "anthropic",
      "cis": 91.3,
      "arena_elo": 1397,
      "gpqa": 74.9,
      "swe_bench": 72.0,
      "math_500": 96.4,
      "humaneval": 92.0,
      "updated_at": "2026-03-14T00:00:00Z"
    }
  ]
}

Agent Economics

GET /agents/costs

Returns AI agent development and inference cost data across different agent types and complexity tiers.

Response

{
  "data": [
    {
      "agent_type": "customer_support",
      "dev_cost_range": { "min": 15000, "max": 50000 },
      "monthly_inference": { "min": 200, "max": 2000 },
      "roi_months": 3,
      "avg_savings_pct": 65
    }
  ]
}

Model Index

GET /models

Returns the full index of all tracked models with provider, pricing, and benchmark data in a single response.

Response

{
  "data": [
    {
      "id": "gpt-4o",
      "provider": "openai",
      "pricing": { "input_1m": 2.50, "output_1m": 10.00 },
      "benchmarks": { "cis": 87.2, "arena_elo": 1381 },
      "context_window": 128000,
      "released": "2024-05-13"
    }
  ],
  "meta": { "total": 22 }
}

API Plans

Plan	Requests/Month	Rate Limit	Price
Free	-	-	No API access
Pro	-	-	No API access
Team	10,000	60/min	$79/user/month
Enterprise	Unlimited	Custom	Custom

Error Codes

Code	Meaning
401	Missing or invalid API key
403	Plan does not include API access
429	Rate limit exceeded
404	Endpoint or resource not found
500	Internal server error

All errors return JSON with a message field:

{
  "error": true,
  "code": 429,
  "message": "Rate limit exceeded. Try again in 42 seconds."
}

Ready to integrate Perffeco data?

Get API access with a Team plan. Start building cost-aware AI applications today.

View Plans