Graph of Thoughts

Real problem-solving combines insights from different approaches. GoT lets an agent explore parallel ideas and then merge compatible partial solutions, outperforming ToT on tasks requiring integration of multiple perspectives.

GoT is a research team where members work on different angles, then merge findings into a unified report — rather than one person following a single outline.

Scenario

Write a business plan: branch A researches market, branch B researches competitors, branch C drafts financials. Aggregate node merges all three into a coherent document.

Solution

In Agent Design Patterns, apply Graph of Thoughts to this scenario: Write a business plan: branch A researches market, branch B researches competitors, branch C drafts financials. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Graph of Thoughts (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Graph of Thoughts happens in the code.

1# Graph of Thoughts — minimal example2from openai import OpenAI3
4client = OpenAI()  # create API client5
6# Ask the model to explain this topic7response = client.chat.completions.create(  # core API call for Graph of Thoughts8    model="gpt-4o-mini",9    messages=[10        {"role": "system", "content": "You explain graph of thoughts clearly."},11        {"role": "user", "content": f"What is graph of thoughts?"},12    ],13    temperature=0,14)15print(response.choices[0].message.content)  # show output for debugging

• Treating Graph of Thoughts as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for graph of thoughts

Quick recap — the most important points from this module.