# Multi-Model Research Analysis

> Multi-Model Research Analysis is a paid API for AI agents from api.ai-open-models.com, paid per call via x402, $0.18958/call, status unknown (last checked 2026-10-31).

Runs a single live-web research task through 2–7 open-weight LLMs simultaneously, returning separate structured analyses with citations, latency, and usage stats per model — no synthesis or ranking.

## Facts

- Endpoint: POST https://api.ai-open-models.com/v1/research/compare?utm_source=zero.xyz
- Price: $0.18958/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-31
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/multi-model-research-analysis-bb8bf936
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_edSlQJlibJoP7rjx8bj9o

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 multi-model-research-analysis-bb8bf936 -d '<json body>'
```

Example prompt: Run this research task — 'What are the latest funding rounds in the AI infrastructure space?' — through Qwen/Qwen3.8-27B, deepseek-ai/DeepSeek-V4.1-Flash, and moonshotai/kimi-k3, pulling 7 web sources, category 'news', and give me each model's independent cited analysis without combining them.

## When to prefer this

Choose this endpoint when you need independent, unmerged perspectives from multiple open-weight models on the same live-web research task — especially for model benchmarking, bias comparison, or when you explicitly do not want a synthesized or ranked answer. Prefer this over the sibling Multi-Model Research Synthesis endpoint when preserving each model's distinct reasoning matters more than a unified conclusion. It is ideal for researchers, analysts, and agent pipelines that consume per-model outputs downstream.

## Known failure modes

- Fewer than 2 models selected — request rejected with validation error
- Invalid or unsupported model identifier in the models array — returns error listing valid enum values
- Objective field empty or exceeds 4000 characters — schema validation failure
- num_results outside 1–10 range — rejected with bounds error
- Web retrieval returns no results for given domain filters or date range — models receive empty evidence context, analyses may be low-quality
- Payment failure via x402 protocol — request not processed
- Model timeout or inference error for one or more models — partial results or error flag per model
- Conflicting include_domains and exclude_domains causing empty retrieval pool

## How this service works

Run one live-web research task through 2–7 buyer-selected open-weight models. Every model receives the same retrieved evidence and returns a separate structured analysis with citations, latency, usage, and validation results. No ranking or synthesis.

## Output

The agent receives one structured response object per selected model, each containing: the model's full text analysis of the research objective grounded in retrieved web evidence, inline citations pointing to the retrieved sources, token usage counts, wall-clock latency for that model, and validation results. No cross-model ranking or synthesized answer is produced — every model's output stands independently.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "models": {
   "type": "array",
   "items": {
    "enum": [
     "moonshotai/kimi-k3",
     "Qwen/Qwen3.6-35B-A3B",
     "zai-org/GLM-5.2",
     "zai-org/GLM-5.3",
     "zai-org/GLM-5.3-Flash",
     "Qwen/Qwen3.8-27B",
     "deepseek-ai/DeepSeek-V4.1-Flash"
    ],
    "type": "string"
   },
   "maxItems": 7,
   "minItems": 2,
   "description": "Required list of 2 to 7 buyer-selected open-weight models. Every model receives the same retrieved evidence.",
   "uniqueItems": true
  },
  "category": {
   "enum": [
    "company",
    "publication",
    "news",
    "personal site",
    "financial report",
    "people"
   ],
   "type": "string",
   "description": "Optional Exa search category."
  },
  "objective": {
   "type": "string",
   "maxLength": 4000,
   "minLength": 1,
   "description": "The live-web research task every selected model must analyze."
  },
  "max_tokens": {
   "type": "integer",
   "default": 2048,
   "maximum": 4096,
   "minimum": 1024,
   "description": "Maximum completion tokens reserved separately for every selected model."
  },
  "num_results": {
   "type": "integer",
   "default": 5,
   "maximum": 10,
   "minimum": 1,
   "description": "Maximum number of ranked web sources to retrieve and read. Each source affects the x402 price."
  },
  "exclude_domains": {
   "type": "array",
   "items": {
    "type": "string",
    "maxLength": 253,
    "minLength": 1
   },
   "maxItems": 10,
   "description": "Optional domain blocklist for web retrieval.",
   "uniqueItems": true
  },
  "include_domains": {
   "type": "array",
   "items": {
    "type": "string",
    "maxLength": 253,
    "minLength": 1
   },
   "maxItems": 10,
   "description": "Optional domain allowlist for web retrieval.",
   "uniqueItems": true
  },
  "end_published_date": {
   "type": "string",
   "maxLength": 64,
   "minLength": 1,
   "description": "Optional latest publication date accepted by Exa."
  },
  "start_published_date": {
   "type": "string",
   "maxLength": 64,
   "minLength": 1,
   "description": "Optional earliest publication date accepted by Exa."
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/multi-model-research-analysis-bb8bf936/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.ai-open-models.com](https://www.zero.xyz/host/api.ai-open-models.com/llms.txt)