Compare · 5 alternatives
Mentat vs Polymarket
The largest USD-volume prediction venue — fast, deep, and centralized at resolution.
Context
Polymarket is the most successful prediction market product of the cycle: pristine UX, deep liquidity on politics, real-time price discovery. Its resolution model is UMA optimistic oracle plus Polymarket Inc. mediation for ambiguous cases. That works until a market is genuinely contested — at which point the centralization is the product. Mentat replaces both layers with cryptographic settlement.
At a glance
| Dimension | Mentat | Polymarket |
|---|---|---|
| Resolution model | zkTLS proof verified on-chain | UMA optimistic oracle + Polymarket Inc. |
| Market creation | Permissionless + AI-drafted + curator-gated | Staff-listed only |
| Chain | Solana | Polygon PoS |
| Fees per trade | Sub-cent network fee + ≤5% protocol cap | ~2% + Polygon gas |
| License | Apache-2.0, self-host | Proprietary |
| Open source | Entire stack | No |
| Settlement latency | Seconds (proof-bound) | Hours to days on disputes |
Resolution
Polymarket relies on UMA's optimistic oracle. Proposers post results, challengers can dispute, holders of UMA token vote on contested outcomes. The system is honest but slow: disputes can take days, and high-profile resolutions (notably several 2024 US election markets) saw vocal community arguments about whether the trigger condition was actually met. Polymarket Inc. retains final discretion to invalidate markets it considers ambiguous. Mentat short-circuits this entire surface with zkTLS proofs against the source-of-truth API. There is no jury vote because there is nothing to vote on.
Market creation
Polymarket markets are staff-listed. Users cannot deploy arbitrary markets. This is a deliberate quality control — and a hard ceiling on the long tail. Mentat permits any user to deploy a market that passes the Validator agent and a curator. The curator backstop ensures quality without staff bottlenecks.
Jurisdiction
Polymarket is restricted in the US (CFTC settlement, ongoing). KYC required in some jurisdictions. Mentat is Solana-native and permissionless at the protocol level; jurisdiction is a front-end concern for any deployment.
Liquidity
Polymarket has the deepest USD liquidity in the space — millions per market on tier-one events. Mentat is pre-mainnet at time of writing. We are not pretending otherwise. The protocol design optimizes for fee economics that work at smaller market sizes (sub-cent Solana fees vs Polygon fees that ate margin on the long tail).
Chain & economics
Polymarket runs on Polygon PoS. Mentat runs on Solana — sub-second confirmations, sub-cent fees, native fast-path for high-frequency trading UI.
When to choose what
Choose Polymarket if you want to trade tier-one US politics with deep liquidity today. Choose Mentat if you are building or operating a prediction market product, want permissionless market creation, and want cryptographic settlement rather than a corporate dispute desk.
FAQ
Has Mentat shipped on mainnet yet?
No — M2 (Creator MVP) is complete, M3 (on-chain) is in active development. Polymarket is live; we are not.
Is UMA settlement actually a problem?
It is a problem when the trigger condition is contestable, which is exactly the case where prediction markets matter most. zkTLS shifts the contest from "what did the data mean" to "did the API actually return X" — and a TLS proof answers that mechanically.
Other comparisons
Mentat vs Kalshi
CFTC-regulated event contracts exchange — US-only, staff-listed markets.
Mentat vs Azuro
Liquidity pool primitive for on-chain bookmakers — not a market protocol itself.
Mentat vs Manifold
Social prediction markets with play-money mana — community forecasting.
Mentat vs Augur
The OG decentralized prediction market on Ethereum — REP-based reporter disputes.