Best AI model release trackers for AI investors in 2026

Q: How many models does LLM Stats track?

LLM Stats tracks 408 canonical models. Its LLM Stats Score combines several categories of model evidence into one comparable number, using an uncertainty-aware approach. It also publishes independent rankings of GPT, Claude, Gemini, Llama, DeepSeek and 300+ AI models. Use these figures to compare models, and check whether a ranking says anything about timing before you rely on it.

Compare Model Odds, Polymarket, Kalshi, Manifold and LLM Stats for tracking AI model releases, with questions to ask before you pick.

By SupaKeeper · 2026-10-23

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

Pick by the question you need answered: when a release might happen, how models compare, or whether a market exists on it. Model Odds ranks Anthropic and OpenAI model families by their estimated chance of a release within 30 days. ^[1]

Option	What it does	Price as stated
Model Odds	Ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days. ^[1]	Free, and so is its API, with no account needed. ^[1]
Polymarket	Describes itself as the world's largest prediction market. ^[2]	not published
Kalshi	Its page calls it a prediction market for trading the future. ^[3] It offers perpetual futures markets through Kalshi Pro. ^[3]	not published
Manifold	Lets users browse markets in categories including politics, technology, sports, culture, business and fun. ^[4]	New users are given 1,000 Mana to start trading. ^[4]
LLM Stats	Tracks 408 canonical models. ^[5]	not published

What AI investors need from AI model release trackers

AI investors usually want to know a few things about a model release. How likely is a model family to ship soon, how was that estimate made, and which past launches does it rest on.

Those are different questions, and they need different tools. A probability tool answers the first. The method and the launch record answer the second and third, and they matter because a number with no record behind it is hard to check. Prediction markets and ranking sites answer adjacent questions, so they may help with a different part of the decision.

The limits are practical. An odds figure is a guess about timing, not an announcement, and a release can land outside the window it was estimated for. Check whether a tool shows how it reached its number, which launches it counted, and whether it can be tested against past dates. The sections below say what each source states and nothing more.

Model Odds for AI investors

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 get their own page that lists every past release. ^[1]

The estimate fits a process to the gaps between recent launches, and it takes into account how long the current wait has already lasted. ^[1] It runs three estimators side by side and shows how far apart they are as an error bar. ^[1] 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]

Signals picked up from company developer kits and docs, 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 gets every company's launches at the same time. ^[1] 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 public API serves odds for each model line, the launch record and service health. ^[1]

Polymarket for AI investors

Polymarket describes itself as the world's largest prediction market. ^[2] Trading is blocked in the United States on the main Polymarket site, which sends US users to a separate US site. ^[2] Polymarket US is a CFTC-regulated Designated Contract Market, where CFTC is the U.S. Commodity Futures Trading Commission. ^[2] Check which site applies to you, and whether it lists a market on the release question you care about.

Kalshi for AI investors

Kalshi's page calls it a prediction market for trading the future. ^[3] It offers perpetual futures markets through Kalshi Pro. ^[3] Check whether a market on a model release exists there before you rely on one.

Manifold for AI investors

Manifold is operated by Manifold Markets, Inc. ^[4] It lets users browse markets in categories including politics, technology, sports, culture, business and fun. ^[4] New users are given 1,000 Mana to start trading. ^[4] Search its technology category for release questions, and check how each market is worded before you rely on it.

LLM Stats for AI investors

LLM Stats publishes independent rankings of GPT, Claude, Gemini, Llama, DeepSeek and 300+ AI models. ^[5] It tracks 408 canonical models. ^[5] Its LLM Stats Score combines several categories of model evidence into one comparable number using an uncertainty-aware approach. ^[5] Use it to compare models, and check whether its ranking says anything about timing before you rely on it.

Matching the tool to the job

To estimate when a model family might ship, Model Odds ranks Anthropic and OpenAI families by their estimated chance of a release within 30 days. ^[1] To check the record behind that number, it lists each past launch with its date and can replay earlier dates. ^[1] ^[1]

To compare models in published rankings, LLM Stats publishes independent rankings of GPT, Claude, Gemini, Llama, DeepSeek and 300+ AI models. ^[5]

To look at whether a market exists on a release question, Kalshi's page describes it as a prediction market for trading the future. ^[3] Polymarket describes itself as the world's largest prediction market. ^[2] Manifold lets users browse markets in technology and other categories. ^[4]

Questions to ask before you choose

Which decision are you making: a timing estimate, a model comparison, or a market you can trade?

Which companies' models matter to your decision, and does each tool list them? Check before you rely on a tool.

Do you need to see how an estimate was built, and whether it can be tested against past dates?

Do you need alerts when a new release lands, and do you want them for every company at once?

Can you use a given market where you live? Check each site's notice for your country.

What does each tool cost for your use? Check each page's price before you commit.

Frequently asked questions

Does Model Odds estimate when Anthropic or OpenAI will release a model?

Model Odds ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days. ^[1] Each figure is a chance over a window, not a fixed date, so read it as a probability. ^[1] 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]

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] You can sign up with a confirmed email to get alerts, and every subscriber gets every company's launches at the same time. ^[1]

Can Polymarket be used for trading from the United States?

Trading is blocked in the United States on the main Polymarket site, which sends US users to a separate US site. ^[2] Polymarket US is a CFTC-regulated Designated Contract Market, where CFTC is the U.S. Commodity Futures Trading Commission. ^[2] Check which site applies to you and whether it lists the market you want before you plan around it.

How many models does LLM Stats track?

LLM Stats tracks 408 canonical models. ^[5] Its LLM Stats Score combines several categories of model evidence into one comparable number, using an uncertainty-aware approach. ^[5] It also publishes independent rankings of GPT, Claude, Gemini, Llama, DeepSeek and 300+ AI models. ^[5] Use these figures to compare models, and check whether a ranking says anything about timing before you rely on it.

Can Model Odds be checked against past dates?

The estimator can replay any past date without using data from after that date. ^[1] It runs three estimators side by side and shows how far apart they are as an error bar. ^[1] Signals from company developer kits and docs, and from third-party model catalogues when two of them agree, can only raise the odds. ^[1]

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

Try Model Odds