agent-to-agent benchmarking

StackRank

The agent-to-agent benchmarking API.
Call us, get a ranked list back.

POST/api/compare
→
ranked JSON in milliseconds
— live output preview
#1
Cohere
0.836
score
#2
OpenAI
0.726
score
#3
Anthropic
0.712
score
— how to call it
request
$ curl -X POST https://stackrank.nanocorp.app/api/compare \
  -H "Content-Type: application/json" \
  -d '{
    "candidates": ["OpenAI", "Anthropic", "Cohere"],
    "criteria": ["price", "latency", "reliability"],
    "weights": {
      "price": 0.4,
      "latency": 0.4,
      "reliability": 0.2
    }
  }'▋
response …
{
  "ranked": [
    {
      "name": "Cohere",
      "score": 0.836,
      "breakdown": {
        "price": 0.82,
        "latency": 0.85,
        "reliability": 0.84
      }
    },
    {
      "name": "OpenAI",
      "score": 0.726,
      "breakdown": {
        "price": 0.55,
        "latency": 0.70,
        "reliability": 0.92
      }
    },
    {
      "name": "Anthropic",
      "score": 0.712,
      "breakdown": {
        "price": 0.50,
        "latency": 0.68,
        "reliability": 0.91
      }
    }
  ],
  "methodology": "Weighted scoring across criteria.",
  "meta": {
    "candidates_evaluated": 3,
    "criteria_used": ["price", "latency", "reliability"]
  }
}
— capabilities
◈
Objective Rankings
Deterministic, reproducible scores based on real benchmark data. No opinions — just numbers.
◇
Any Criteria
Pass any criteria and custom weights. Price vs latency vs reliability — you decide the tradeoffs.
◉
Agent-Native
Designed to be called by other AI agents. Structured JSON in, ranked leaderboard out.
— request schema
POST /api/compare
{
  "candidates": string[]    # required
  "criteria":   string[]    # required
  "weights":    Record<string, number>  # optional
}