How AI investors gauge market expectations for model launches

Learn how AI investors can gauge market expectations for model launches, what to prepare, and how to compare tools that estimate release odds.

By SupaKeeper · 2026-10-24

SupaKeeper is published by LunarWerx, which makes some of the tools compared here.

To gauge market expectations for an AI model launch, write down a dated question, record estimates from more than one source, compare them with official company posts, and note how each estimate changes over time. Model Odds is one tool that publishes estimates of this kind for Anthropic and OpenAI ^[1]

Option	What it does	What to check
Model Odds	Ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[1] It is free, and so is its API, with no account needed ^[1]	Check how each past launch is recorded and whether the estimate matches the time window in your question.
Official company posts	See the company's own pages for what it covers.	Check which companies and model families the posts address, and how each launch is dated.
Independent AI news coverage	See the outlet's own pages for what it covers.	Check which companies and model families it reports on, and whether a launch is confirmed by an official source.

What you need before you start

Have a clear list of the companies and model families you want to follow. Anthropic and OpenAI are the two companies that the Model Odds ranking covers, so a list that includes either one lets you compare their estimates directly ^[1]

Write down the question you are answering. A useful question names one model family and a time window, such as whether a release is likely in the next month. A vague question makes almost any estimate look right.

Set up a simple dated record for your estimates. Note the source, the date you read it, the model family, and the estimate. Later, you can check whether a source's view moved before or after an official post appeared.

Keep the official posts from the companies you follow within reach. A launch is measured against what the companies themselves publish, so you will need those posts when you review any estimate.

Decide how you will receive updates, whether by checking a page on a set schedule or by signing up for email alerts. Settle this first so a change in the estimate does not slip past you.

Step by step

Write the question with a model family and a time window. Keep the same wording for every source so the answers can be compared.

Collect the estimates from each source you use, and record the date each one was taken. An estimate from last week and one from today may not describe the same situation.

Check how each source built its launch history. Look for whether past launches are dated, and whether each one was confirmed by an official post.

Compare the estimates side by side. If they disagree widely, treat the gap as a measure of uncertainty rather than picking the number you prefer. Write down the reason for any large gap, since that reason is the most useful thing to review later.

Look at the signals behind any estimate. Note whether the evidence comes from the companies' own materials or from outside reports, and whether two independent sources point the same way.

Check how the estimate handles time that has already passed. A long wait since the last launch can change the picture, so note whether the source accounts for that.

Once a launch happens or does not happen, compare each estimate against what the companies publish. Record the result next to the estimate.

Review your method after each launch or missed launch. Adjust the question, the sources, or the record so the next comparison is fairer.

Tools that can do it

Model Odds is made by the team that publishes this site. It ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[1] Anthropic and OpenAI each have their own page that lists every past release ^[1] The estimate is a model of the gaps between recent launches that also counts how long the current wait has already lasted ^[1] It runs three estimators side by side and publishes the gap between them as the error bar ^[1] Signals from company developer kits and documentation, and from third-party model catalogues when two of them agree, can only raise the odds ^[1] The estimator can replay any past date without using data from after that date ^[1] You can sign up with a confirmed email to get alerts, and every subscriber receives every company's drops at the same time ^[1] Model Odds is a website with a public API ^[1] The website and the API share one snapshot that is cached for up to five minutes ^[1]

The other options are official company posts and independent AI news coverage. Read them with the same questions you use for Model Odds: which companies and model families they cover, how each launch is dated, and whether a launch is confirmed. Their own pages describe what each one covers, so check those pages before relying on either.

Common mistakes to avoid

Treating an estimate as a date. An estimate is a chance over a window, so write it down that way and do not plan around one exact day.

Comparing estimates from different days. An estimate taken before a major post and one taken after it may describe different situations. Keep the dates together with the numbers.

Averaging away disagreement. When sources differ widely, the gap is the useful part. Write down the reason for the gap instead of hiding it.

Skipping the launch record. An estimate is only as useful as the history behind it. Check how past launches were dated and confirmed before you compare anything.

Using a signal without checking it. One report is not the same as two sources that agree. Note how many independent sources point the same way.

Forgetting the scope. A source may cover only some companies or model families. Write the scope next to every estimate so the comparison stays fair.

Frequently asked questions

How do AI investors gauge market expectations for model launches?

They usually write down a dated question for a model family, collect estimates from more than one source, and compare those estimates with official company posts once a launch happens or does not happen. The gap between estimates is useful, because it shows how uncertain the picture is. Keeping a dated record of what each source said lets you see later which views moved and when. The steps matter more than any single tool, so the same method works with whichever sources you choose.

What should be checked before relying on a model launch estimate?

Check how past launches were recorded and whether each one was confirmed by an official post. Check the date of the estimate, because an old one may not match current conditions. Check which companies and model families the source covers, since a source may cover only some of them. Finally, check whether the source explains its method clearly enough for you to judge it.

Is Model Odds free to use?

Model Odds is free, and so is its API, with no account needed ^[1] It is a web app that runs in the browser and needs JavaScript ^[1] Its API serves odds for each model line, the launch record, and service health ^[1] The website and the API share one snapshot that is cached for up to five minutes ^[1]

Does Model Odds rank launches from companies other than Anthropic and OpenAI?

Model Odds ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[1] Each of those two companies has its own page that lists every past release ^[1] For other companies, check the tool's own pages to see which companies and model families it lists before you rely on it.

How does Model Odds decide whether a past launch counts?

Each past launch it uses has to be confirmed by an official post and an independent timeline, and it is listed with its date ^[1] The estimate also uses signals from company developer kits and documentation, and from third-party model catalogues when two of them agree. Those signals can only raise the odds ^[1]

Model Odds is free, and so is its API, with no account needed. ^[1]

Try Model Odds

Sources

Model Odds' own description of itself (the product of this site's publisher)