Overview

LLMs tuned on our data have outperformed frontier models

Forecasting

Get probability estimates on binary forecasting questions with Lightning Rod's hosted models. No training required—call the API and receive calibrated forecasts.

Foresight-v3

foresight-v3 is Lightning Rod's latest forecasting model. Pass any binary forecasting question and receive a probability estimate between 0 and 1. Use it via the OpenAI-compatible API.

Setup

Sign up at dashboard.lightningrod.ai to get your API key and $50 of free credits.

pip install openai lightningrod-ai

from openai import OpenAI

client = OpenAI(
    api_key="lightningrod-api-key", 
    base_url="https://api.lightningrod.ai/api/public/v1/openai"
)

Get forecasts

Pass any forecasting question to foresight-v3.

questions = [
    "Will the Fed cut rates by 25bp in March 2026?",
    "Will Apple announce a new Vision Pro model in 2026?",
    "Will the S&P 500 close above 6000 by end of 2026?",
]

for q in questions:
    response = client.chat.completions.create(
        model="LightningRodLabs/foresight-v3",
        messages=[
            {"role": "system", "content": "Answer as a probability between 0 and 1."},
            {"role": "user", "content": q}
        ]
    )
    print(f"Q: {q}")
    print(f"A: {response.choices[0].message.content}\n")

API reference

See the API docs for full details on parameters (temperature, max_tokens, etc.).

Try it

Run the foresight-v3 notebook in Google Colab for a complete walkthrough.