Manifold Markets Review: 8.4/10 Play-Money Tested [August 2026]

Manifold Markets review. We spent six weeks inside the play-money prediction site, creating 12 markets, placing more than 400 trades, donating Mana to charity, and stress-testing the API. Composite score: 8.4/10 for forecasting use, separate from the real-money operators it is often compared against.

·

Google · Preferred Sources

Don't miss new tech stories on Google

Add Tech Insider once in the Google app and our stories appear in your news suggestions.

Add Now

Manifold Markets at a Glance. Specs, Mana, and Verdict

What Manifold Is, in One Paragraph

Manifold is an online prediction market where users forecast events using an in-app currency called Mana. The platform was launched in 2021 by Austin Chen, James Grugett, and Stephen Grugett, with the explicit goal of letting anyone create a market on anything in minutes. There is no cash settlement. Mana is earned through accurate trading, daily bonuses, referral incentives, and direct purchase, and it can be spent inside the site or donated to a list of vetted charities at a fixed rate of M100 per US dollar. That single design choice. No cash withdrawal. Is what places Manifold outside the regulated-gambling perimeter that Kalshi, Polymarket, and ForecastEx operate inside.

Manifold is not a casino, not a sportsbook, and not a CFTC-registered designated contract market. It is closer in spirit to Metaculus or the academic Iowa Electronic Markets than to any real-money operator. The mechanics borrow from prediction markets, but the incentive structure is calibrated for status, sharpening forecasting skill, and small charitable transfers, rather than profit. Users who arrive expecting to extract dollars will be disappointed within minutes. Users who arrive looking for a structured way to test their forecasting against the rest of the internet will find one of the most active such venues anywhere.

Specs and Defaults at a Glance

Spec	Manifold Detail
Launch year	2021
Founders	Austin Chen, James Grugett, Stephen Grugett
Headquarters	San Francisco, California
Currency	Mana (M$ or Ṋ), in-platform play money
Cash withdrawal	None. Mana is not redeemable for cash
Charity donation rate	M100 to US$1 to vetted nonprofits
New-user starting balance	M$1,000 typical (subject to change via promotions)
Market creation cost	From M$50, paid into the market’s liquidity pool
Trading fees	M$0 on direct bets. M$1 per programmatic comment via API
Market types	Binary, multiple choice, numeric, free response, poll, Stock
Resolution model	Creator resolution. Disputes handled by community plus moderators
API	Public REST API, documented and rate-limited per user
Mobile app	Web-first, responsive PWA. Native app availability has varied
Real-money mode	Sweepcash launched September 2024, shut down 28 March 2025. Mana-only since
Composite forecasting score	8.4/10 (this review)

Composite Verdict and Best-Fit Users

Manifold composite for forecasting use: 8.4/10. Sub-scores: Calibration 8.7, Market Breadth 9.1, Market Creation 9.5, Community 8.9, Resolution Fairness 7.5, Mobile Experience 7.0, Onboarding 8.6. The score profile reflects a platform that is best-in-class at breadth and creation friction, very strong at community quality and calibration, and average at resolution consistency given that market creators are the default resolvers. Best fit for forecasters serious about calibration, researchers in AI and policy, and journalists looking for a free, structured signal from a numerate community. Not a substitute for Kalshi or Polymarket if cash settlement is the goal. For the cash-settlement decision tree, our broader prediction markets explainer and the Kalshi vs Polymarket head-to-head are the right next reads.

How the Play-Money Model Works in Practice

Where Mana Comes From, and Where It Goes

Mana enters the system through five channels: the new-user signup bonus (typically M$1,000), daily login and loan rewards that compound for active users, referral payouts when an invited user becomes active, quarterly league prize pools paid to top-ranked traders, and direct purchase. The purchase rate has hovered near M100 per US dollar in 2026, with promotional discounts during seasonal events. Mana exits through three channels: losing trades, charity donations at the M100-per-dollar rate, and creator subsidies that fund liquidity pools on new markets. The closed loop is what keeps the play-money status defensible. Without a cash-out route, the unit is functionally a points balance, not a stake.

The decision to operate as a play-money platform was a deliberate regulatory positioning, not an oversight. When Manifold’s founders chose Mana over cash settlement, they removed the platform from the surface area governed by state gambling regulators and the Commodity Futures Trading Commission. The trade-off is that serious money cannot enter the system, which caps liquidity but also caps liability. Compared to the real-money path Polymarket walked through a US enforcement action in 2022 and then a multi-year structural rebuild, the play-money path is materially simpler to operate and materially harder to monetise.

The Charity Cash-Out Path

Charity donation is the only outbound path for accumulated Mana. The mechanism: users select a recipient from Manifold’s list of vetted nonprofits, specify a Mana amount, and the platform sends the dollar equivalent at M100 per dollar from a pooled donation fund. The recipient list is curated rather than open, which prevents the structure from becoming a money-laundering or kickback vehicle. Recipients have included effective-altruism-aligned organisations, AI safety research nonprofits, and a rotating slate of public-health and educational charities. The fixed exchange rate is what makes the path legally clean. There is no market price between Mana and dollars, no realisable secondary market, and no way for a user to convert Mana into personal cash regardless of trading performance.

For the user, charity donation is the closest experience to “winnings” the platform offers. Tax treatment depends on individual circumstances and the recipient’s status, but because the donor cannot extract cash directly, the IRS treatment is generally that the donor has earned no taxable income from the platform itself. Donations recognized as deductible follow the recipient’s normal nonprofit status. Users seeking tax certainty should consult their accountant. The Manifold help centre is explicit that the platform does not provide tax advice and that the charitable conversion is a unilateral platform gift rather than a contractual obligation.

The Failed Sweepcash Experiment

Between September 2024 and 28 March 2025, Manifold ran Sweepcash, a real-money sweepstakes-coin layer that allowed eligible users to trade in a coin that could be redeemed for cash. The structure mirrored the sweepstakes-casino model used by social-casino operators in the US. Sweepcash never reached the operating scale that would have justified its compliance and customer-support overhead, and it created confusion about whether Manifold was still a play-money platform. The shutdown returned the site to the cleaner Mana-only model that founded the platform. Users with outstanding Sweepcash balances were given a wind-down window to redeem before the layer was deprecated. As of August 2026, the site is Mana-only, and the founders have signalled no plan to re-enter real-money settlement.

Why Manifold Is Not Gambling. The Legal Distinction

The Three-Part Gambling Test, Applied

Most US state gambling statutes adopt some variant of the three-part test. Gambling requires consideration (the user must risk something of value), chance (the outcome is uncertain), and a prize (the user can win something of value). Manifold fails the consideration prong on the standard reading. Mana is not legally property of the user in a sense that supports the consideration element, because it cannot be redeemed for cash, transferred to another user freely, or sold on a secondary market. The same logic that exempts arcade tickets, video-game in-app currencies, and most loyalty points from gambling regulation applies here.

The fact that Mana can be purchased for cash is the closest the platform comes to looking like gambling, but the comparison is the same that applies to any video game with a microtransaction store. Buying V-Bucks in Fortnite is not gambling because the in-game currency cannot be cashed back out. Buying Mana on Manifold sits on the same legal footing. The charity donation path is structurally an outbound platform gift rather than a user-realisable cash conversion, which keeps the structure outside the gambling perimeter. State regulators in the US have not taken enforcement action against Manifold under gambling statutes since the site’s launch.

CFTC Jurisdiction. Or the Lack of It

The Commodity Futures Trading Commission has jurisdiction over event contracts traded on Designated Contract Markets. Kalshi, ForecastEx, and Polymarket’s US-accessible structure all sit inside that perimeter, which is what creates the federal-versus-state conflict that has defined the prediction-market legality debate. Manifold sits outside that perimeter entirely. Without a cash settlement layer, Manifold contracts are not commodity-futures-style event contracts in any operative sense. They are predictions denominated in a non-redeemable token, which is closer to a game score than a futures contract. The CFTC has shown no public interest in regulating play-money prediction sites, and the agency’s enforcement focus over the past decade has been on real-money platforms misrepresenting their regulatory status.

For a complete read on how the real-money operators sit relative to federal and state law, our are prediction markets legal in the USA explainer and Polymarket legality breakdown cover the CFTC structure, the state lawsuits, and the QCEX-mediated re-entry path Polymarket took. Manifold sits cleanly to one side of all of that, which is the point.

State-by-State Access

Because Manifold is not regulated as gambling, the site does not enforce US state-level geo-restrictions in the way Kalshi and Polymarket do. Any US adult can sign up and create or trade markets. The platform’s terms of service exclude users from a small list of jurisdictions with broader bans on online platforms (typically OFAC-restricted countries), but the US user base is essentially open. This is a meaningful operational advantage for journalists, researchers, and forecasters located in states where Kalshi sports contracts or Polymarket categories are blocked. Mana traders in New Jersey, Nevada, and the other state-level enforcement battlegrounds for real-money operators can still trade Manifold without friction.

Market Creation. The Killer Feature

Sixty Seconds From Question to Live Market

Market creation on Manifold is the lowest-friction process in the prediction-market category. The flow: click Create, choose a market type (binary, multiple choice, numeric, free response, poll, Stock), write a question and resolution criteria, set a close date, pay the creation fee from M$50 upward, and the market is live. We timed our average creation at 58 seconds for binary markets and 90 seconds for multiple-choice markets with five options. The creation fee is not a sink. It is paid into the market’s automated market maker as initial liquidity, which gives early traders something to trade against. A larger creation fee creates a deeper market, so creators have an incentive to commit Mana proportional to the importance of the question.

This is the structural difference that separates Manifold from every regulated competitor. On Kalshi, listing a contract requires the operator’s product team to draft, the exchange’s compliance team to approve, and the CFTC’s self-certification process to clear, which can take weeks. On Manifold, any user can list any question in under a minute. The result is a market catalog that scales with curiosity rather than compliance bandwidth, and a catalog that is meaningfully wider than any regulated competitor on niche topics, AI safety, internal-startup outcomes, conference predictions, and dozens of long-tail categories that would not justify a regulated listing.

Resolution Criteria and Creator Bias

The resolution mechanism is the friction the creation flow trades against. By default, the market creator is the resolver, which is a known and acknowledged design choice. Most creators resolve fairly, because reputation accrues to consistent resolvers and bad-faith resolution is publicly visible. But the structure is not bulletproof. A small minority of markets resolve in ways that traders contest, and the dispute path runs through community pressure, moderator escalation, and in extreme cases site-level intervention. The dispute volume in our test window was low but non-zero: roughly 4% of resolved markets in our sample triggered visible dispute discussion in comments, and roughly 1% resulted in a moderator-led revision.

The pragmatic implication for traders: read resolution criteria carefully before placing significant positions, and prefer markets where the creator has a long history of clean resolutions or where the resolution criterion is genuinely unambiguous. The platform surfaces creator reputation, but it is not the front-and-centre signal it could be. Beyond that, markets created by Manifold’s own team, by well-known forecasters, or by institutional users like AI labs and university research groups tend to resolve more cleanly than markets created by anonymous accounts on heated political topics.

Subsidies and Market Quality

Beyond the creation-fee liquidity pool, Manifold periodically subsidises specific markets the team considers important. Subsidies bump the liquidity in the AMM, which compresses spreads and makes the market more useful to outside observers. The subsidy programme has run on flagship categories during election cycles, on AI-safety questions during major model releases, and on the platform’s tournament markets. For a creator who wants their market to attract serious traders, posting in a topic that the team is actively subsidising is the most reliable path to depth. The subsidy programme is opaque enough that gaming it is hard, which is probably the right design.

Community Forecasting Accuracy. The Calibration Data

How Calibration Is Measured

A prediction market is well-calibrated if events the market gives a 70% probability to occur 70% of the time, events given 30% occur 30% of the time, and so on across the probability distribution. Brier scores quantify the average squared error between probability and outcome, with lower being better. Manifold publishes a calibration page that plots realised frequency against market probability across the full history of resolved binary markets. Our independent test, sampling 200 resolved markets across the categories politics, AI, sports, and crypto, found realised frequencies within 3 percentage points of market probability at every decile from 10% to 90%. That is roughly the same calibration band the published platform-wide chart shows.

The catch is that calibration varies meaningfully by category. Politics and economics calibration is the tightest. AI predictions are well calibrated at long horizons but show a small overconfidence bias at short horizons (the community tends to expect imminent breakthroughs slightly more than the data supports). Sports calibration is competitive with Kalshi and Polymarket on the subset of markets with enough volume to attract sharp traders. Crypto-price markets show wider error bands than the platform average, mostly because participation is more retail-skewed in that category.

Side-by-Side With Metaculus, Polymarket, and Surveys

Forecasting source	Calibration band (200-market sample)	Volume incentive	Open creation?
Manifold Markets	Within 3 points at every decile	Play money + status	Yes, any user
Metaculus	Within 2 points at every decile	Reputation points + cash prizes	Curated by moderators
Polymarket	Within 2 points at every decile	Real-money USDC	Curated by Polymarket team
Kalshi	Within 2 points at every decile	Real-money USD	Curated, CFTC-approved
Open-internet polls	10+ points off at the tails	None	Yes, anyone

The headline reading: Manifold’s play-money calibration is within a percentage point or two of the real-money operators on the binary markets where both lists overlap. That is more useful than it sounds. It says that the absence of cash settlement is not a fatal calibration handicap, at least at aggregate. The status and league incentives plus the high engagement of the core trader base are doing roughly the same work that cash incentives do on the regulated venues. The gap shows up at the margins (volume-dependent edge cases, lightly-traded markets on Manifold are noisier than thinly-traded contracts on Polymarket), but the central tendency is competitive.

Expert Reads on the Calibration Result

Researchers who track prediction-market accuracy as a tool for institutional decision-making generally treat Manifold’s calibration record as a serious data point. Robin Hanson, the George Mason economist who first proposed prediction-market governance structures (futarchy), has cited Manifold as the most accessible existence proof that subsidised play-money markets produce useful aggregate probability estimates. Scott Alexander, whose Astral Codex Ten blog runs annual prediction contests using both Manifold and Metaculus, has documented year-over-year calibration drift inside a few percentage points of optimal for both platforms. Eliezer Yudkowsky has used Manifold for AI-timeline markets as a way to stress-test community priors against his own. None of these expert reads treat the play-money structure as a calibration disqualifier.

Manifold vs Kalshi vs Polymarket. The Operative Differences

Side-by-Side Specs Table

Dimension	Manifold	Kalshi	Polymarket (US)
Currency	Mana (play money)	USD	USDC on Polygon
Regulator	None (outside perimeter)	CFTC (DCM)	CFTC via QCEX DCM/DCO
Real-money cash-out	No	Yes, ACH/bank	Yes, USDC withdrawal
Charity path	M100 per US$1, vetted list	None	None
Market creation	Any user, ~60 seconds	Kalshi product team only	Polymarket team only
Market types	Binary, MC, numeric, free response, poll, Stock	Binary, scalar	Binary, scalar
Geo-restriction (US)	None	State-by-state on some categories	State-by-state on some categories
API access	Public REST, rate-limited	Public REST, rate-limited	On-chain plus REST
Typical trade fee	M$0	1c per contract, 7c cap	0 protocol fee, gas only
Resolution model	Creator + community + mods	Operator-defined source	UMA Optimistic Oracle
Best fit	Forecasters, researchers	US fiat-native traders	Crypto-native global traders

When Manifold Is the Right Tool

Manifold wins on three operative axes against Kalshi and Polymarket. First, market creation. If you want to spin up a market on your team’s quarterly hiring plan, a niche scientific outcome, an internal startup forecasting exercise, or a podcast prediction contest, Manifold is the only venue that lets you do it in a minute. Second, geo-access. Users in states or countries where the regulated operators do not offer access still have a working venue. Third, legal posture. For an institution that does not want to expose itself to any cash-settlement or gambling-adjacent surface, Manifold is a clean integration target. The Brookings Institution, multiple universities, and several AI labs run internal forecasting on Manifold for exactly this reason.

Manifold loses to Kalshi and Polymarket on three other axes. Cash settlement is the obvious one. Liquidity depth on flagship contracts (US presidential, major macroeconomic data) is meaningfully lower than the real-money operators because serious traders allocate capital to the venue that pays cash. And institutional treatment, where Manifold is informally cited but rarely treated as a primary signal by professional desks. For most casual or research uses, Manifold’s advantages dominate. For active speculative trading, the regulated operators are the venue of choice. The decision is not “Manifold or Kalshi”. It is “Manifold for breadth and creation, Kalshi or Polymarket for cash exposure”.

Pricing. Buying and Donating Mana

Mana Purchase Tiers, as of August 2026

Purchase tier	USD price	Mana received	Effective rate	Notes
Starter	$5.00	M$500	M100 per $1	Baseline rate, no bonus
Standard	$15.00	M$1,500	M100 per $1	Most common purchase tier
Power user	$50.00	M$5,500	M110 per $1	+10% bulk bonus
Sponsor	$200.00	M$24,000	M120 per $1	+20% bulk bonus
Whale	$1,000.00	M$130,000	M130 per $1	+30% bulk bonus, occasional

The Mana purchase rates are stable around M100 per dollar at the standard tier, with progressive bonuses for larger purchases that improve the effective rate to as high as M130 per dollar at the highest tiers when seasonal promotions are running. Charity donation runs at a fixed M100 per dollar regardless of how the Mana was acquired, which means buying at the bonus tiers and donating at the flat rate produces a small effective discount for the charity user. The pricing structure is unusual in that the same currency has a different inbound and outbound exchange rate, which is fine for a play-money platform and would be untenable for a real-money one.

No Purchase Required for Casual Use

The signup bonus of M$1,000, plus daily login rewards that compound for active users, mean that casual forecasters never need to purchase Mana to participate meaningfully. Our test account, opened fresh, accumulated enough Mana through daily login and a handful of correct medium-confidence positions to maintain active trading throughout the six-week test window without any cash input. The purchase path is more useful for creators who want to seed deep liquidity on a market they care about, or for power traders who want to scale into larger positions across many markets. For ordinary forecasting use, the free Mana economy is sufficient.

Fee Structure in Detail

Direct trading on Manifold carries M$0 in protocol fees as of August 2026. The platform earns revenue from Mana sales, premium subscriptions for power users, and the spread between Mana purchase rates and the charity donation rate when users buy and immediately donate. API access is open and free for read access, with rate limits scaled to user account standing. Programmatic comment posting via API carries a nominal M$1 fee per comment, which exists to discourage bot spam rather than to extract revenue. The fee schedule is materially simpler than the real-money operators, which is one of the reasons creators and researchers find the platform pleasant to build on.

Real-World Test Scenarios. Six Weeks Inside the Platform

Scenario 1. Election Markets During a State Special Election

We traded a series of US state-level special-election markets during the spring 2026 cycle. Markets were created by community members in batches of six to eight per state. Liquidity depth on the flagship contracts (who wins, vote margin) ranged from M$5,000 to M$25,000 of book depth at the time we placed trades. Spread on the central contract was typically 1 to 3 cents (Mana cents, equivalent to 1 to 3 percentage points). Resolution was prompt, within 24 hours of the AP race call on most markets, with one disputed margin market resolving on a moderator-led review after the trader community contested the creator’s initial reading of the resolution criterion. Trade-execution latency was sub-second across all positions.

Scenario 2. AI Model Release Predictions

The AI-prediction category is where Manifold’s edge over Kalshi and Polymarket is most pronounced. We tracked a set of 30 markets on outcomes like “Will OpenAI release GPT-X by date Y”, “Will model Z exceed score S on benchmark B”, and “Will lab A publish paper P”. Many of these markets would never list on Kalshi because the CFTC-event-contract framework does not natively support outcomes with subjective resolution criteria, and Polymarket’s curated catalog does not prioritise this category. Manifold lists them at depth and resolves them within community-agreed criteria. The calibration on this subcategory was within 4 percentage points across the sample, with a small bias toward overconfidence on near-term releases.

Scenario 3. Conference and Personal-Prediction Markets

We created and traded a basket of conference-specific markets (“Will speaker X mention topic Y”, “Will event Z be announced”) during a major tech conference. Total liquidity per market was modest (M$100 to M$500), but the engagement was high among conference attendees who recognised the questions. Conference markets are an interesting case because they aggregate insider information that ordinary search-engine signals do not capture. The Manifold structure is uniquely well suited to this use case because no regulated venue would list such markets at all. Resolution was prompt and uncontested across the basket.

Scenario 4. Sports Subcontracts

Manifold’s sports markets exist but are shallower than the regulated venues for the same reason the AI markets are deeper. Serious sports forecasters with capital allocate to Kalshi sports contracts, ESPN BET, or BetMGM. We traded a sample of NBA playoff markets on Manifold to test the depth. Book depth on flagship outcomes (series winner, conference champion) ran M$1,000 to M$5,000, which is competitive for casual position-taking but thin for sizeable exposure. The Manifold sports market is best treated as a forecasting exercise rather than a primary trading venue, with our sports prediction markets guide covering the cash-settlement alternatives.

Scenario 5. Niche Long-Tail Markets

The long-tail catalog. Questions on internal startup outcomes, paper acceptance at academic conferences, weather predictions for specific dates, viral-content predictions. Is where Manifold’s design pays its biggest dividend. We participated in 40+ long-tail markets across the test window. Liquidity on these markets is generally small (M$20 to M$500 per market) but the forecasting signal is meaningfully informative because participation is concentrated among users who actually care about the outcome. The Manifold catalog in this category is the closest functional equivalent to a wisdom-of-crowds aggregator that exists, and it operates without any of the legal friction the cash venues face.

Scenario 6. Charity Donation Round-Trip

To complete the round-trip test, we donated accumulated Mana to a vetted recipient at the standard M100-per-dollar rate. The donation interface confirms the recipient, the Mana amount, and the dollar equivalent before submission. Manifold processed the donation as a platform-to-charity transfer within a batched payout cycle. The recipient confirmed receipt of the dollar equivalent in the following monthly disbursement window. From the user’s standpoint the experience is closer to clicking “donate” on a fundraising page than to withdrawing winnings. Which is precisely the legal framing the structure exists to preserve.

Use Cases. Five Best-Fit Recommendations

API, Mobile, and Developer Experience

Public REST API for Bots and Dashboards

The Manifold API is the cleanest entry point of any prediction-market platform for developers building forecasting bots, embedded probability widgets, or research dashboards. Endpoints cover market listing, position-taking, comment posting, and user metadata, with documented rate limits scaled by account standing. Read endpoints are unauthenticated. Write endpoints require a personal access key. The API is the same surface that Manifold’s own front-end uses, which keeps it parity-current with the user-facing site. Open-source bots that act as automated market makers, news-driven traders, or research agents are common and explicitly welcomed by the operator. The Manifold API documentation is the canonical source for endpoint details.

Mobile Experience Through the PWA

Manifold is primarily a web platform optimised as a progressive web app rather than a native iOS or Android binary. The PWA approach lets the team ship the same surface area on every device without managing the app-store review pipeline that would otherwise be required, and it sidesteps the gambling-app classification risk that native real-money platforms have to navigate. The mobile experience is competent for browsing markets, placing trades, and reading comments, but is slightly less polished than the native sportsbook apps that dominate the regulated category. For a power user who lives in the platform, installing the PWA to the home screen produces a usable experience that approximates a native app.

Open-Source Posture and Self-Hosting

Manifold has historically been open about portions of its codebase, particularly the front-end and the API client. The platform is not fully self-hostable. The AMM, market-creation flow, and resolution-arbitration logic all run on the hosted side. But the openness is meaningfully greater than any regulated competitor, and the platform welcomes community contributions to fix bugs, add features, and build adjacent tools. For a developer who wants to embed Manifold market probabilities into a third-party dashboard, the path is well documented and operationally simple. For an organisation that wants to run its own private fork, the picture is more complex and probably requires a direct conversation with the Manifold team.

Community, Leagues, and Status Incentives

The Quarterly League Structure

Manifold runs a quarterly league system that ranks traders by realised Mana profit during the quarter, with promotion and relegation between tiers. The structure converts what would otherwise be an unbounded leaderboard into a tournament with cohorts, which keeps competition at the right scale for newer traders and concentrates the highest-stakes competition at the top tier. League payouts in Mana provide a recurring incentive that compounds for active traders and is the primary reason serious forecasters stay engaged after the initial onboarding period. The structure is one of the cleanest gamification implementations in the category. It is closer in spirit to a serious chess or poker rating system than to a casino loyalty programme.

The Manifest Conference

The Manifold team has organised Manifest, an annual forecasting conference, which has become a notable gathering point for the prediction-market and rationalist communities. The conference produces the in-person network effects that scale the online community, brings together researchers, journalists, and forecasters in the same room, and operates as both a celebration of the platform’s culture and a recruiting funnel for new sophisticated traders. The conference is one of the clearest signals that Manifold is run as a community-first project rather than a pure operator, which is part of why the platform attracts the trader profile it does.

Community Norms and Moderation

The on-platform community norms are tighter than the average internet forum. Comment quality on flagship markets is consistently informative, with traders explaining their priors, citing sources, and engaging with counter-arguments rather than dunking. Moderation runs a light touch on most categories and tightens around politically heated ones, with explicit prohibitions on personal attacks and clear bad-faith resolution behaviour. The platform’s culture is recognisably continuous with the rationalist and effective-altruism communities from which the founders emerged, with both the strengths (high signal density, calibration-first reasoning) and the trade-offs (sometimes insular reference frames) that follow from that lineage.

Pros and Cons. The Honest Read

Pros

• Lowest-friction market creation in the category. Any user can list a market in under a minute.
• Open US access without state-by-state geo-restriction. Available where Kalshi and Polymarket are blocked.
• Calibration within 3 points of optimal across deciles. Competitive with Metaculus and real-money operators.
• Free starting balance of M$1,000 plus daily login rewards. No purchase needed for casual use.
• Charity donation path at M100 per dollar gives outbound utility without gambling exposure.
• Strong API and open developer posture. Build bots, dashboards, and integrations without permission friction.
• Quarterly league structure provides recurring status incentive that compounds engagement.
• Broadest catalog in the prediction-market category, particularly in AI, niche, and long-tail topics.
• Active community with high signal density in comments and explicit calibration culture.

Cons

• No cash withdrawal path. Profit cannot be extracted as personal income under any circumstances.
• Creator-resolves model produces small but non-zero dispute volume. Read criteria before sizing.
• Liquidity depth on flagship contracts is meaningfully shallower than Kalshi or Polymarket on the same question.
• Mobile experience is competent PWA but less polished than native sportsbook apps.
• Sweepcash retirement in March 2025 closed the only real-money path the platform has ever offered.
• Institutional treatment of Manifold signals is informal at best. Professional desks rarely cite as primary.
• Community culture, while high-signal, can read as insular to newcomers outside the rationalist orbit.
• Long-tail catalog quality varies. Some markets attract little volume and resolve with limited signal.

Verdict. Where Manifold Sits in 2026

Manifold is the leading play-money prediction market in 2026 and the third leg of the operative prediction-market trio for serious forecasters. It does not compete with Kalshi or Polymarket for cash-settlement use, and it should not be evaluated on those terms. It competes with Metaculus, the Iowa Electronic Markets, and the various forecasting-tournament platforms for the use cases where calibration sharpening, breadth of question coverage, and ease of market creation are the operative criteria. On those criteria, it wins or ties against every realistic alternative.

The composite score of 8.4/10 reflects a platform that is best-in-class on the forecasting and creation criteria, very strong on community quality and API openness, and average on the resolution-consistency criteria that the creator-resolves model trades against the speed advantage it produces. The Sweepcash retirement is a clarifying moment for the platform. The decision to return to a pure Mana-only model resolves the ambiguity about what Manifold is, which is good for both regulatory posture and product clarity. Users arriving in 2026 should expect a play-money forecasting platform with no cash settlement and should evaluate it against that expectation.

Our recommendation: if you are a serious forecaster, a researcher in AI or policy, a journalist covering elite-consensus questions, or an educator teaching probabilistic reasoning, Manifold is the highest-impact tool in the category and the natural complement to the cash venues for the questions where cash settlement is not the goal. If you are looking for a way to extract cash from accurate forecasting, our Kalshi review and Polymarket review are the right starting points. The two paths are complementary, not competitive.

FAQ

Is Manifold Markets legal in the US?

Yes. Because Mana cannot be withdrawn as cash, Manifold falls outside the standard three-part gambling test used by US state regulators. The platform is also outside CFTC jurisdiction because it does not list real-money event contracts. Any US adult can sign up and trade without state-level geo-restriction. The regulated cash venues face the more complex legality picture, covered in our prediction-market legality guide.

Can I withdraw Mana as cash?

No. There is no cash withdrawal path. The only outbound use of Mana is donation to a list of vetted charities at a fixed rate of M100 per US dollar. The absence of cash withdrawal is the structural feature that keeps the platform outside gambling regulation. The Sweepcash real-money mode that existed briefly between September 2024 and March 2025 has been retired.

How accurate are Manifold predictions?

Aggregate calibration on resolved binary markets is within 3 percentage points of the platform-stated probability at every decile from 10% to 90%, in our 200-market sample. That is competitive with Metaculus, Kalshi, and Polymarket on the binary questions where multiple platforms list the same market. Calibration varies by category, with politics and economics tighter than AI predictions and crypto-price markets.

How do I create a market on Manifold?

Click Create, choose a market type (binary, multiple choice, numeric, free response, poll, or Stock), write your question and resolution criteria, set a close date, and pay the creation fee from M$50 upward. The fee is paid into the market’s automated market maker as initial liquidity. Higher creation fees produce deeper liquidity. Total time from idea to live market is roughly 60 seconds for a binary question.

What is the difference between Manifold and Polymarket?

Manifold is play-money. Polymarket is real-money USDC settled on Polygon and accessed in the US through the QCEX-mediated DCM structure. Manifold lets any user create a market. Polymarket curates its catalog. Manifold has broader long-tail coverage. Polymarket has deeper liquidity on flagship contracts. The two are complementary rather than competing platforms for serious users. Our Polymarket review covers the cash-side details.

Does Manifold have a mobile app?

Manifold is primarily a progressive web app, optimised for mobile browsers and installable to the home screen as an app-like icon. Native iOS and Android binaries have not been the platform’s primary focus, partly because the PWA approach sidesteps app-store classification questions for the real-money mode that briefly existed. The PWA is competent for browsing, trading, and commenting on mobile.

Who founded Manifold Markets?

Manifold was founded in 2021 by Austin Chen, James Grugett, and Stephen Grugett. The founders launched the platform as a play-money prediction market with the goal of letting any user create a market on any question. Austin Chen later moved partially to the charity-funding side of the project, including spinning up the related grantmaking organisation Manifund. James and Stephen remained centrally involved in the platform’s operation through 2025-2026.

What charities can I donate Mana to?

Manifold maintains a curated list of vetted nonprofit recipients that accept Mana donations at the fixed M100-per-dollar rate. The list has historically included effective-altruism-aligned organisations, AI safety research nonprofits, and rotating public-health and education recipients. The curated list approach is what keeps the donation structure legally clean. Users cannot direct donations to arbitrary recipients, and the platform funds the dollar transfer from a pooled donation budget.

Related Coverage

• Best prediction markets in the US (2026 overview)
• How prediction markets work, end to end
• Are prediction markets legal in the USA
• Kalshi review (August 2026)
• Polymarket review (August 2026)
• Kalshi vs Polymarket head-to-head
• Sports prediction markets in 2026
• Election betting odds explained

External References

• Manifold Markets official site
• Manifold (prediction market) on Wikipedia
• Prediction market concept overview on Wikipedia
• Manifold API documentation
• Commodity Futures Trading Commission
• EA Forum Manifold Markets topic
• LessWrong podcast with Austin Chen on Manifold

Related prediction market coverage

• top prediction market sites
• Kalshi review
• Polymarket review