Sandwiching bots:

These bots construct a “sandwich” around another transaction by combining frontrunning and backrunning (the inverse of frontrunning). Assume, for instance, that a bot notices a sizable buy order for a token on a DEX. If so, it may attempt to purchase the same token before the execution of the initial transaction (frontrunning), and then sell it back after another bot or person purchases it at a higher price (backrunning).

Frontrunning bots:

These bots keep monitoring the mempool for significant or lucrative transactions, then attempt to carry out their own transactions ahead of them. Assume, for instance, that a bot notices a sizable buy order for a token on a DEX. In that situation, it may attempt to purchase the same token before the initial transaction and then sell it back at a greater cost after the initial order increases the token’s demand and price.

Flash loan bots:

These bots carry out intricate trades needing significant sums of capital via flash loans, which are brief loans returned inside the same block. For instance, a bot can borrow a significant sum of ETH via a flash loan, exchange it for another token on a DEX, exchange it back for ETH on another DEX, and pay back the loan with interest all in one block. If done properly, the bot might make money from this.

Arbitrage bots: 

These bots search for price variations between the same asset on various marketplaces and carry out trades to take advantage of them. For instance, an arbitrage bot can acquire a token on Uniswap and then sell it on Sushiswap for a profit if the token is less expensive on Uniswap than on Sushiswap.