# AI Model Benchmark Evidence Search for Agent Builders

> AI Model Benchmark Evidence Search for Agent Builders is a paid API for AI agents from market.datapackvibe.com, paid per call via x402, $0.01/call, status unknown (last checked 2026-10-21).

Searches live public web pages for AI model benchmark evidence tailored to agent builders, returning source links and snippets.

## Facts

- Endpoint: POST https://market.datapackvibe.com/x402/pulse-web-ai-model-benchmarks-agent-builders?utm_source=zero.xyz
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-21
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/ai-model-benchmark-evidence-search-for-agent-builders-818f2f79
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_6fRY9BTbYo-qL5TyMw-kp

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 ai-model-benchmark-evidence-search-for-agent-builders-818f2f79 -d '<json body>'
```

Example prompt: Can you search for the latest AI model benchmark evidence relevant to agent builders — things like reproducible results and methodology datasets — and give me up to 5 source links with snippets?

## When to prefer this

Choose this endpoint when you need live, sourced web evidence about AI model benchmarks specifically relevant to agent builders — not static datasets or pre-indexed databases. Ideal when you want fresh snippets and links rather than pre-computed model cards or internal evaluations.

## Known failure modes

- Query returns no results if topic is too niche or no recent pages are indexed
- max_results exceeding 10 is rejected by schema validation
- Slow or unavailable upstream web search causes timeout or empty response
- Snippets may be outdated if the live web crawl hasn't recently indexed the target pages

## How this service works

AI model benchmark evidence for Agent builders: Search live public web pages about ai model benchmark evidence, focused on agent builders. Returns source links and snippets, not endorsements.

## Output

Returns a list of up to 10 web page snippets and their source URLs drawn from live public web pages about AI model benchmarks, focused on agent builder contexts. Results are informational only and not endorsements.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "query": {
   "type": "string",
   "default": "AI model benchmarks methodology datasets reproducible results Agent builders",
   "description": "Optional focused query."
  },
  "max_results": {
   "type": "integer",
   "default": 5,
   "maximum": 10,
   "minimum": 1
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/ai-model-benchmark-evidence-search-for-agent-builders-818f2f79/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from market.datapackvibe.com](https://www.zero.xyz/host/market.datapackvibe.com/llms.txt)