All prompt templates
ResearchQuality

Investment research

When to use

High-stakes research where a single model’s confident hallucination is expensive. Fusion’s cross-review flags weak claims.

Recommended setup

Panel: GPT-5.5, Claude Opus 4.8, Gemini 3.1 ProJudge: Claude Opus 4.8

The prompt

Act as an investment analyst researching {COMPANY / ASSET}. Summarize the business model, market, competitive moat, financial health, and key risks. Lay out the bull case and the bear case explicitly. Be honest about uncertainty and what data is missing. This is analysis for research purposes, not financial advice.

API example

{
  "model": "openrouter/fusion",
  "messages": [
    {
      "role": "user",
      "content": "Act as an investment analyst researching {COMPANY / ASSET}. Summarize the business model, market, competitive moat, financial health, and key risks. Lay out the bull case and the bear case explicitly. Be honest about uncertainty and what data is missing. This is analysis for research purposes, not financial advice."
    }
  ]
}

A Fusion call bills as the panel + judge combined — estimate it with the cost calculator. how to call the Fusion API.