There is a great comparison in the Quant Whitepaper. Compares each aspect of Interoperability/performance and covers all the current Crypto Interop. Projects. I recommend checking that out.

Cosmos is receiving a lot of hype and getting fired up after their massive launch. In reality, their Interop. solution isn't out of testing yet.

They have a similar inter-chain/zone hub solution, as do a lot of other Interoperability projects. These middleman Blockchains make integration and overhead complexity much more of an issue.

This article here, written by the Quant team helps to outline some of the fundamental differences: https://medium.com/@gverdian/how-overledger-differs-from-interoperability-blockchains-97fff3094ff8

A Telegram member (Seq) recently had a discussion about Cosmos in the Quant Telegram. I will quote his explanations. Insightful.

"Cosmos are a long way away from getting any interoperability, all thats been released as main net is the cosmos hub and staking, you can't connect any zones to it as their interoperability protocol IBC is still in research / prototype phase. One of the main differences between cosmos and polkadot is that cosmos wants every App to have their own blockchain, which stores the state of the app and the security is up to the App. So they have to get a load of people to deploy nodes for every single app made. Whereas Polkadot uses shared security from the hub validators and uses a small randomly selected validators of the token hub. The problem for Cosmos imo is that how secure all these chains going to be? Just look at the amount of useless coins there are now and your supposed to get 100's of people running nodes for the app. Ultimately its going to result in the hub being very secure and the zones that connect to the hub not being very secure and centralised and easy to take control with POS. Now many think that it won't matter because the hub is what connects them so doesn't matter about the security of the other chains, but thats not the case. The hub doesn't keep track of the state of each of the App blockchains. The hub only keeps track of transfers between blockchains. To prevent double spending (swapping them in one chain and then swapping the same tokens in another chain etc). The problem is though it means you have to trust the app blockchain to trade with it. As you don't know whether there has been some bug and someone has stolen a load of tokens, or the chain has been compromised before you exchange some of your tokens for theirs. So all its going to result in, is exactly the same as the Internet now, where you don't trust the majority of apps, but the large big vendors with lots of validators like BTC / ETH etc you do trust. And from an investment point of view. The ATOM token is only used for staking for nodes within the Cosmos hub only. There will be many hubs and they will each have their own token (to try and incentivise people to run nodes). The ATOM token is designed to be hyperinflationary so the price falls over time per token as they have a minimum inflation rate of 7% and a maximum of 20%. Meaning that they will keep releasing tokens. They don't want ATOM to be used for trading, they want it to be used for staking. The incentivisation isn't from the price of ATOM increasing but the fact that for each transaction that goes through the hub you receive a % of the transaction fee. The transaction fee can be paid in a variety of currencies - BTC, ETH, Dash and many many more. The other problem that Cosmos and Polkadot have is being able to connect to chains that aren't finite with their consensus, so for chains like BTC aren't compatible at all with Polkadot without btc forking - from the Polkadot FAQ on their website:

Can Polkadot connect any blockchain?
Polkadot can connect any previously existing blockchain if it matches two criteria:

It must have the ability to form compact and fast light-client proofs over the finality and validity of its blocks and state change information (this would include new UTXOs in a Bitcoin-like chain or logs in an Ethereum-like chain).
There must be a means by which a large set of independent authorities (perhaps up to one thousand) can authorise a transaction. This could include recognition of threshold signatures, such as the Schnorr scheme, or a smart contract able to structure logic against a multi-signature condition.
Bitcoin and Bitcoin-like chains fall short on these characteristics. To address the first criteria, Polkadot validators can simply run a full Bitcoin node. To address the second criteria, either a soft-fork allowing extra-protocol controls over funds or a hard-fork enabling a threshold-signature-friendly signing scheme such as Schnorr is needed. Neither are impossible goals, however a significant degree coordination would be required to achieve them.

Both Cosmos and Polkadot use Peg Zones for other chains like ETH which involves adding another blockchain where each of the validators of the pegzone also have to run a ETH node. It then involves signing the transaction and trying to make it compatible witht he other chain. Basically its really difficult to do, they have to have a separate Pegzone for every External chain (like any other blockchain you know of now) adding more overhead, transaction fees, restrictions in being able to connect to certain blockchains."