Skip to content
Term Finance logo

Term Finance Review

Fixed-rate loans with published fees and a worked example

of 10

Term Finance scores 5.2 in DeFi protocols. Its fee pages now give rates and a worked example, but its contracts carry a non-commercial licence, no audit is listed after 2024, and its vaults lost funds to a governance exploit in August 2026. Desk research checked November 6, 2026.

Overall score

5.2 = (8 + 2 + 6 + 6 + 4) / 5

Borrower fees

Servicing typically 0.30%-0.50%/yr; liquidated damages typically 8%

Licence

CC BY-NC-ND 4.0 (term-finance-contracts)

Audits

Six linked reports, four firms, Apr 2023 to Oct 2024

Upgrade control

Term Foundation initiates; TERM holders can veto via Protocol Veto Governor

Incident

Aug 23, 2026 vault governance exploit; direct lending contracts unaffected

Scores

What it does & fees
8.0
Licence & code openness
2.0
Audit disclosure
6.0
Backing & transparency
6.0
Governance & control
4.0

Each criterion carries equal weight. The overall score is the average of these 5 scores — (8.0 + 2.0 + 6.0 + 6.0 + 4.0) ÷ 5 = 5.2.

Pros

  • Fees 8: servicing fee typically 0.30% to 0.50% a year, with a worked example
  • Liquidated damages stated: typically 8% of debt covered, 5.2% liquidator and 2.8% protocol
  • Five dated, linked core audits from Sigma Prime, Runtime Verification and Dedaub
  • Loans overcollateralized on-chain; example margins 150% initial, 125% maintenance

Cons

  • Licence 2: contracts are CC BY-NC-ND 4.0, no commercial use or derivatives
  • No audit listed for Term V2, Terminal 1 or the vaults; the latest core audit is November 2023
  • USDC and ETH vaults hit by a governance exploit on August 23, 2026; root cause unpublished
  • Veto window, quorum and signer details for upgrades are not stated

How Term Finance scores in DeFi protocols

This entry rates Term Finance, a non-custodial Ethereum protocol for fixed-rate, fixed-term loans against crypto collateral, in our DeFi protocols ranking. Rates are set in sealed-bid auctions, and since August 2026 Term V2 also lets borrowers and lenders post or take orders at their own rate. The basis is desk research of Term's documentation, its blog, its terms of use and its GitHub organization, checked on November 6, 2026. No loan was taken or offered. The category has five criteria, each marked 0 to 10 and weighted equally, as set out in our ratings methodology.

CriterionWhat the sources showMark
What it does & feesFunction stated plainly. Borrowers pay a servicing fee of typically 0.30% to 0.50% a year, deducted up front, with a worked example. Liquidated damages are typically 8% of debt covered (5.2% liquidator, 2.8% protocol). Lenders pay no protocol fee.8
Licence & code opennessterm-finance-contracts carries Creative Commons BY-NC-ND 4.0: readable, but no commercial use and no derivatives.2
Audit disclosureSix linked reports from four firms: five on the core protocol (Sigma Prime, Runtime Verification, Dedaub, April to November 2023) and one on Blue Sheets (ChainLight, October 2024). No audit is listed for Term V2, Terminal 1 or the vaults. No bug bounty is stated.6
Backing & transparencyLoans are overcollateralized in protocol contracts and marked to market with Chainlink feeds; example margins are 150% initial and 125% maintenance. Shortfalls are shared as a haircut by a loan's lenders.6
Governance & controlThe Term Foundation initiates upgrades and access-control changes; TERM holders can veto them through a Protocol Veto Governor. Veto periods, thresholds and signers are not stated. The USDC and ETH vaults lost funds to a governance exploit on August 23, 2026.4

The overall score is the plain average: (8 + 2 + 6 + 6 + 4) / 5 = 26 / 5 = 5.2, up from 5.0 at the July 24, 2026 check. Three marks moved. What it does & fees rose from 5 to 8 because the fee pages now give the rates and a worked example; in July the amounts did not render. Audit disclosure fell from 9 to 6. The July review credited six audit firms; the audits page lists four, the latest core review is from November 2023, and nothing is listed for the V2 contract changes or Terminal 1 that went live in 2026. Governance & control rose from 3 to 4 under the governance scale applied across the category on November 6, 2026: a documented controller with no stated delay scores 4.

Strongest criterion

Fees stated with rates and a worked example (8)

Term charges borrowers two things. The servicing fee is "typically ranging from 0.30%-0.50% per annum", applied to the loan amount on an actual/360 basis and taken from the proceeds. The docs work an example: a 28-day, 100,000 USDC loan at 0.5% costs about 38.89 USDC. Liquidated damages are "typically set at 8%" of debt covered in a liquidation, split 5.2% to the liquidator and 2.8% to the protocol. The example states there are no fees to lenders, though borrowers price the fee into their bids. The mark is not higher because both figures are given as typical levels rather than fixed rates, and the V2 order-markets announcement states no fees of its own.

Weakest criteria

A non-commercial, no-derivatives licence (2): a deficiency

The core contracts repository carries Creative Commons Attribution-NonCommercial-NoDerivatives 4.0. That is not a software licence approved as open source, and it is stricter than the Business Source License: it never converts, and it bars both commercial reuse and modified versions. The category gives 2 for a licence that states non-commercial or no-derivatives terms. Derive and Synthetix score 0, because no public source exists for their live contracts.

Governance and control (4): a deficiency with a live incident

A Term blog post from April 2025 describes the split. The Term Foundation operates the protocol and initiates actions under the DEVOPS role, which covers "smart contract upgrades or changes to access controls"; TERM holders can veto them through a Protocol Veto Governor, and DAO fund moves through a Token Veto Governor. The pages opened give no veto window, quorum, multisig composition or timelock. On August 23, 2026 an attacker "targeted the governance of the affected vaults", Term's USDC and ETH vaults. Term's October 1 update says its direct fixed-rate lending contracts were not compromised. For the USDC pool it reports $987,095.52 recovered against $1,638,604.34 of recorded losses, and it says no stolen ETH has been recovered. Term has not published a root cause or a change to its controls on the pages opened. The category gives 4 where the controller is documented but no timelock or delay is stated, as for dHEDGE; a documented timelock with a bypass would score 5. This criterion scores that disclosure, so the exploit is reported here rather than scored, and the missing veto window and thresholds keep the mark at 4.

Audits that stop at 2023 for the core (6): partly a deficiency

The five core audits are dated and linked: Sigma Prime reviewed releases 0.1.0, 0.2.0 and 0.6.0, Runtime Verification checked the clearing algorithm, and Dedaub reviewed 0.4.0. ChainLight reviewed Blue Sheets in October 2024. The page lists nothing after that, although Term V2 changed the core contracts in August 2026 and Terminal 1 was announced in May 2026. That is a disclosure gap, not a finding that they went unreviewed. The mark sits below Derive's 7, which has a named audit of its current core.

Backing: on-chain collateral with known failure modes (6)

Borrowers lock collateral in the protocol when they bid, and positions below maintenance margin can be liquidated at once, with no margin call. If liquidations do not cover a loan, the docs say its Term Repo Tokens are redeemed at a haircut shared by all its lenders. Term's own blog documents an April 2025 tETH oracle error that caused unintended liquidations for 18 users, who were reimbursed. The mark stays at 6: the design is overcollateralized and on-chain, but no reserve or collateral dashboard was found and both incidents touched how collateral is valued or held. Our guide to how DeFi lending works explains margin ratios and liquidations.

What the score does not cover

The score measures disclosure, not returns or the quality of any vault curator. Term Repos are non-callable: borrowers cannot repay early and lenders cannot redeem before maturity, though Repo Tokens can be sold. A borrower who misses the repurchase window, typically 24 hours, defaults. Recovery for vault depositors depends on outstanding Repo Tokens repaying, and distributions were scheduled for October 2 and December 15, 2026. Our checklist for reading a DeFi lending market covers what to check before lending.

Term Finance vs Maple Finance and Derive

Maple Finance, the other fixed-rate lender rated here, scores 7.0. The two tie on fees (8). Maple leads on licence (5 against 2), audits (10 against 6), backing (7 against 6) and governance (5 against 4), because its audit archive runs to October 2026 and its timelock is documented. Derive also scores 5.2; it is listed first only because ties are ordered alphabetically. Term leads on licence (2 against 0), because Derive has published no source for its live contracts, and on backing (6 against 5). Derive leads on fees (9 against 8), audits (7 against 6) and governance (5 against 4).

Who it suits

On these marks, Term Finance suits borrowers and lenders who want a fixed rate for a fixed term, set by auction or order, and who can hold a position to maturity. It suits less well depositors who need open-source code, recent audits of the current release, or published governance parameters. Governance and audits are the criteria most likely to move at the next check.

How it compares

Sources

Scored from published primary sources rather than hands-on use. Every figure above comes from one of the pages below, on the date shown. Manufacturers change prices and specifications without notice — if something here no longer matches the source, tell us and we will correct it.

  1. 1.Term Finance docs — servicing fee — checked
  2. 2.Term Finance docs — liquidated damages — checked
  3. 3.Term Finance docs — security audits — checked
  4. 4.GitHub — term-finance-contracts LICENSE (CC BY-NC-ND 4.0) — checked
  5. 5.Term blog — Staking and governance on Term (15 Apr 2025) — checked
  6. 6.Term blog — Term Vault recovery: distribution update (1 Oct 2026) — checked
  7. 7.Term blog — Term V2 (4 Aug 2026) — checked
  8. 8.Term blog — tETH oracle incident postmortem (28 Apr 2025) — checked
  9. 9.Term Finance docs — maturity, rollovers and liquidation — checked
  10. 10.Term Finance docs — Term Repo risks — checked
  11. 11.Term Finance docs — terms of use (16 Jun 2026) — checked

Frequently asked questions

Why does Term Finance score 5.2?

It is the plain average of five equal marks: fees 8, licence 2, audits 6, backing 6 and governance 4. Published fees carry the score; the restrictive licence and weak governance disclosure hold it back.

What changed since the July 2026 score?

The overall score rose from 5.0 to 5.2. Fees rose from 5 to 8 because the fee rates are now readable, and audit disclosure fell from 9 to 6 because the audits page lists four firms, not six, and nothing for the 2026 releases. Governance rose from 3 to 4 when one governance scale was applied across the category on November 6, 2026.

Did the August 2026 exploit affect Term's lending contracts?

Term says its direct fixed-rate borrowing and lending contracts were not compromised. The attack targeted the governance of its USDC and ETH vaults. The governance mark scores what Term discloses about control; the exploit is reported alongside it, and Term has not published a root cause.

What would raise Term Finance's score?

Listing audits of Term V2 and Terminal 1, publishing veto windows and signer thresholds, and a root-cause report on the vault exploit would address the audit and governance marks. A permissive licence would lift the licence mark.

How does Term Finance compare with Maple Finance?

Maple Finance scores 7.0. The two tie on fees; Maple leads on licence, audits, backing and governance, mainly through its audit archive to October 2026 and a documented timelock.

Nothing here is financial advice. Editorial policy · How we score · How we're funded.