Turn Claude into an Accurate Forecaster

Give Claude the ability to produce calibrated forecasts on any question about the future, one question or a hundred at once. Forecast every country's GDP growth, every company's next earnings, every candidate's election odds, or any set of questions you can put in a spreadsheet. Three modes are available:

Binary: "Will X happen?" → calibrated probability (0 to 100)
Numeric: "What will the price/value/count be?" → percentile distribution (p10 through p90)
Date: "When will X happen?" → percentile dates (p10 through p90, as YYYY-MM-DD)

Effort is configurable. Default effort resolves automatically: HIGH for a single forecast, LOW for many. HIGH is ~$1.20 per row; LOW is ~$0.09 to $0.20 per row.

Here, we forecast when Anthropic and OpenAI will IPO and what their first-day valuations will be.

Add FutureSearch to Claude Code if you haven't already:

claude mcp add futuresearch --scope project --transport http https://mcp.futuresearch.ai/mcp

Now Claude can forecast anything. For binary questions:

Will Anthropic IPO before July 2027? Forecast the probability.

Claude calls FutureSearch's forecast MCP tool:

Tool: futuresearch_forecast
├─ data: [{"question": "Will Anthropic IPO before July 1, 2027?",
│          "resolution_criteria": "Resolves YES if Anthropic common shares trade on a public exchange before July 1, 2027."}]
└─ forecast_type: "binary"

→ Submitted: 1 row for probability forecasting.
  Session: https://futuresearch.ai/sessions/...

Tool: futuresearch_progress
→ Running: 0/1 complete, 1 running (30s elapsed)

...

Tool: futuresearch_progress
→ Completed: 1/1 (0 failed) in 270s.

Tool: futuresearch_results
→ probability: 42, rationale: "Anthropic engaged Wilson Sonsini as legal counsel..."

For numeric questions, Claude automatically uses percentile mode:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO,
in billions of USD.

Tool: futuresearch_forecast
├─ data: [{"question": "What will Anthropic's first-day market cap be?", ...},
│         {"question": "What will OpenAI's first-day market cap be?", ...}]
├─ forecast_type: "numeric"
├─ output_field: "market_cap"
└─ units: "billions USD"

→ Submitted: 2 rows for numeric percentile forecasting.

...

Tool: futuresearch_results
→ Saved 2 rows with market_cap_p10 through market_cap_p90 columns.

Add the FutureSearch connector if you haven't already. Then ask Claude any forecasting question:

Will Anthropic IPO before July 2027? Forecast the probability.

Or for numeric forecasts:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO. Give me the full probability distribution in billions of USD.

Go to futuresearch.ai/app and enter any forecasting question:

Will Anthropic IPO before July 2027? Forecast the probability.

Or for numeric forecasts:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO. Give me the full probability distribution in billions of USD.

pip install futuresearch
export FUTURESEARCH_API_KEY=your_key_here  # Get one at futuresearch.ai/app/api-key

Binary forecasting for yes/no probabilities:

import asyncio
import pandas as pd
from futuresearch.ops import forecast

questions = pd.DataFrame([
    {
        "question": "Will Anthropic IPO before July 1, 2027?",
        "resolution_criteria": "Resolves YES if Anthropic common shares trade on a public exchange before July 1, 2027.",
    },
    {
        "question": "Will OpenAI IPO before January 1, 2028?",
        "resolution_criteria": "Resolves YES if OpenAI common shares trade on a public exchange before January 1, 2028.",
    },
])

async def main():
    result = await forecast(input=questions, forecast_type="binary")
    print(result.data[["question", "probability", "rationale"]])

asyncio.run(main())

Numeric forecasting for percentile distributions:

valuation_questions = pd.DataFrame([
    {
        "question": "What will Anthropic's first-day market cap be when it IPOs?",
        "resolution_criteria": "Market capitalization at close of first trading day, in billions USD.",
        "background": "Anthropic raised $30B at a $380B valuation in early 2026.",
    },
    {
        "question": "What will OpenAI's first-day market cap be when it IPOs?",
        "resolution_criteria": "Market capitalization at close of first trading day, in billions USD.",
        "background": "OpenAI raised at $852B in March 2026. CFO guided toward 2027 listing.",
    },
])

async def main():
    result = await forecast(
        input=valuation_questions,
        forecast_type="numeric",
        output_field="market_cap",
        units="billions USD",
    )
    print(result.data[[
        "question", "market_cap_p10", "market_cap_p25",
        "market_cap_p50", "market_cap_p75", "market_cap_p90",
    ]])

asyncio.run(main())

Date forecasting answers "when will X happen?":

ipo_dates = pd.DataFrame([
    {
        "question": "When will Anthropic IPO?",
        "resolution_criteria": "Date Anthropic common shares first trade on a public exchange.",
    },
])

async def main():
    result = await forecast(
        input=ipo_dates,
        forecast_type="date",
        output_field="ipo_date",
        effort_level="HIGH",
    )
    print(result.data[["question", "ipo_date_p10", "ipo_date_p50", "ipo_date_p90"]])

Date columns are YYYY-MM-DD strings. Forecasters may return the literal "never" for percentiles in the indefinite future.

Use context when all rows share common framing:

result = await forecast(
    input=geopolitics_questions,
    forecast_type="binary",
    context="Focus on EU regulatory and diplomatic sources. Assume resolution by end of 2027.",
)

FutureSearch forecast that Anthropic would list first, around March 2027, with an estimated first-day market cap of $630 billion (ranging from ~$350B to ~$980B). OpenAI would follow in May 2027, with a median first-day cap of $1.0 trillion, though with significant probability it comes in below its $852B private valuation.

Key factors the research agents surfaced:

Timing depends on private capital access. Both companies can raise tens of billions privately, reducing IPO urgency. The forecast gave >10% probability each doesn't IPO within 3 years.
Anthropic gets the bigger IPO pop. Enterprise companies (like Snowflake) tend to see bigger first-day gains than consumer companies (like Uber).
OpenAI valuation depends on ChatGPT sentiment. Revenue is overwhelmingly consumer subscriptions. Unless revenue trends strongly toward their $100B goal, public markets may not tolerate the multiple.

The full analysis is in Anthropic and OpenAI IPO Timelines and Valuations.

Going deeper

Forecast When Anthropic and OpenAI Will IPO: date mode end-to-end.
Forecast Anthropic and OpenAI IPO Valuations: numeric, high effort.

Built with FutureSearch. See the forecast documentation for all parameters and output formats. Related guides: Find Profitable Prediction Market Trades, Forecast Outcomes for a List of Entities, Value a Private Company.

Turn Claude into an Accurate Forecaster

Binary: "Will X happen?" → calibrated probability (0 to 100)

Numeric: "What will the price/value/count be?" → percentile distribution (p10 through p90)

Date: "When will X happen?" → percentile dates (p10 through p90, as YYYY-MM-DD)

Effort is configurable. Default effort resolves automatically: HIGH for a single forecast, LOW for many. HIGH is ~$1.20 per row; LOW is ~$0.09 to $0.20 per row.

Here, we forecast when Anthropic and OpenAI will IPO and what their first-day valuations will be.

Add FutureSearch to Claude Code if you haven't already:

claude mcp add futuresearch --scope project --transport http https://mcp.futuresearch.ai/mcp

Now Claude can forecast anything. For binary questions:

Will Anthropic IPO before July 2027? Forecast the probability.

Claude calls FutureSearch's forecast MCP tool:

Tool: futuresearch_forecast
├─ data: [{"question": "Will Anthropic IPO before July 1, 2027?",
│          "resolution_criteria": "Resolves YES if Anthropic common shares trade on a public exchange before July 1, 2027."}]
└─ forecast_type: "binary"

→ Submitted: 1 row for probability forecasting.
  Session: https://futuresearch.ai/sessions/...

Tool: futuresearch_progress
→ Running: 0/1 complete, 1 running (30s elapsed)

...

Tool: futuresearch_progress
→ Completed: 1/1 (0 failed) in 270s.

Tool: futuresearch_results
→ probability: 42, rationale: "Anthropic engaged Wilson Sonsini as legal counsel..."

For numeric questions, Claude automatically uses percentile mode:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO,
in billions of USD.

Tool: futuresearch_forecast
├─ data: [{"question": "What will Anthropic's first-day market cap be?", ...},
│         {"question": "What will OpenAI's first-day market cap be?", ...}]
├─ forecast_type: "numeric"
├─ output_field: "market_cap"
└─ units: "billions USD"

→ Submitted: 2 rows for numeric percentile forecasting.

...

Tool: futuresearch_results
→ Saved 2 rows with market_cap_p10 through market_cap_p90 columns.

Add the FutureSearch connector if you haven't already. Then ask Claude any forecasting question:

Will Anthropic IPO before July 2027? Forecast the probability.

Or for numeric forecasts:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO. Give me the full probability distribution in billions of USD.

Go to futuresearch.ai/app and enter any forecasting question:

Will Anthropic IPO before July 2027? Forecast the probability.

Or for numeric forecasts:

Forecast the first-day market cap of Anthropic and OpenAI when they IPO. Give me the full probability distribution in billions of USD.

pip install futuresearch
export FUTURESEARCH_API_KEY=your_key_here  # Get one at futuresearch.ai/app/api-key

Binary forecasting for yes/no probabilities:

import asyncio
import pandas as pd
from futuresearch.ops import forecast

questions = pd.DataFrame([
    {
        "question": "Will Anthropic IPO before July 1, 2027?",
        "resolution_criteria": "Resolves YES if Anthropic common shares trade on a public exchange before July 1, 2027.",
    },
    {
        "question": "Will OpenAI IPO before January 1, 2028?",
        "resolution_criteria": "Resolves YES if OpenAI common shares trade on a public exchange before January 1, 2028.",
    },
])

async def main():
    result = await forecast(input=questions, forecast_type="binary")
    print(result.data[["question", "probability", "rationale"]])

asyncio.run(main())

Numeric forecasting for percentile distributions:

valuation_questions = pd.DataFrame([
    {
        "question": "What will Anthropic's first-day market cap be when it IPOs?",
        "resolution_criteria": "Market capitalization at close of first trading day, in billions USD.",
        "background": "Anthropic raised $30B at a $380B valuation in early 2026.",
    },
    {
        "question": "What will OpenAI's first-day market cap be when it IPOs?",
        "resolution_criteria": "Market capitalization at close of first trading day, in billions USD.",
        "background": "OpenAI raised at $852B in March 2026. CFO guided toward 2027 listing.",
    },
])

async def main():
    result = await forecast(
        input=valuation_questions,
        forecast_type="numeric",
        output_field="market_cap",
        units="billions USD",
    )
    print(result.data[[
        "question", "market_cap_p10", "market_cap_p25",
        "market_cap_p50", "market_cap_p75", "market_cap_p90",
    ]])

asyncio.run(main())

Date forecasting answers "when will X happen?":

ipo_dates = pd.DataFrame([
    {
        "question": "When will Anthropic IPO?",
        "resolution_criteria": "Date Anthropic common shares first trade on a public exchange.",
    },
])

async def main():
    result = await forecast(
        input=ipo_dates,
        forecast_type="date",
        output_field="ipo_date",
        effort_level="HIGH",
    )
    print(result.data[["question", "ipo_date_p10", "ipo_date_p50", "ipo_date_p90"]])

Date columns are YYYY-MM-DD strings. Forecasters may return the literal "never" for percentiles in the indefinite future.

Use context when all rows share common framing:

result = await forecast(
    input=geopolitics_questions,
    forecast_type="binary",
    context="Focus on EU regulatory and diplomatic sources. Assume resolution by end of 2027.",
)

Key factors the research agents surfaced:

Timing depends on private capital access. Both companies can raise tens of billions privately, reducing IPO urgency. The forecast gave >10% probability each doesn't IPO within 3 years.

Anthropic gets the bigger IPO pop. Enterprise companies (like Snowflake) tend to see bigger first-day gains than consumer companies (like Uber).

OpenAI valuation depends on ChatGPT sentiment. Revenue is overwhelmingly consumer subscriptions. Unless revenue trends strongly toward their $100B goal, public markets may not tolerate the multiple.