# Capability Benchmark

> Capability Benchmark is a paid API for AI agents from phion.systems, paid per call via x402, $0.004/call, status unknown (last checked 2026-10-18).

Calculates reproducible capability performance scores from versioned, evidence-bound test cases for AI agent services

## Facts

- Endpoint: POST https://phion.systems/v1/paid/trust/capability-benchmark?utm_source=zero.xyz
- Price: $0.004/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-18
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/capability-benchmark-963a81d2
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_S44VJyCeMVVxpkA2fwyBY

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability capability-benchmark-963a81d2 -d '<json body>'
```

Example prompt: Run a capability benchmark on our latest agent service build using versioned evidence-bound test cases and give me a reproducible performance score I can track across versions.

## When to prefer this

Choose this endpoint when you need a deterministic, reproducible performance score for an AI agent capability grounded in versioned evidence — particularly when you need to compare scores across builds, audits, or time periods. Prefer this over ad-hoc evaluation methods when reproducibility and evidence traceability are required, such as in regulated or trust-sensitive agentic deployments.

## Known failure modes

- Malformed or missing input payload returns a validation error
- Unrecognized capability or evidence version identifiers may return a not-found error
- Non-deterministic inputs that cannot be versioned may fail reproducibility checks
- Service unavailability returns a 503 or timeout error
- Insufficient USDC balance (x402 payment failure) blocks the request

## How this service works

Calculate reproducible capability performance from versioned evidence-bound cases

## Output

Returns a reproducible, deterministic capability performance score derived from versioned, evidence-bound test cases. The result is tied to specific evidence versions so scores can be consistently reproduced and compared across time or agent versions.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input",
  "output"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "body"
   ],
   "properties": {
    "body": {
     "type": "object",
     "additionalProperties": true
    }
   }
  },
  "output": {
   "type": "object",
   "required": [
    "example"
   ],
   "properties": {
    "example": {
     "type": "object",
     "additionalProperties": true
    }
   }
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json"
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/capability-benchmark-963a81d2/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from phion.systems](https://www.zero.xyz/host/phion.systems/llms.txt)