As of the creation of this market in late 2023, md5 is widely considered insecure due to readily available collision attacks. However, no first preimage attacks have been performed yet. First preimage attacks are attacks on hash functions that find inputs producing a given output hash where no input was previously known. This market does not include second preimage attacks, which are a subset of collision attacks that have already been performed against md5. A first preimage attack with computation complexity of 2^123.4 has been theorized, but that's not much faster than a naive brute force attack, and it's definitely not practical with current hardware. A successful first preimage attack on md5 by the end of 2027 would likely require either significant advances in md5 attack theory or quantum computers to become viable.

This market will be immediately resolved as "YES" if an md5 preimage for a "nothing-up-my-sleeve number" (see https://en.wikipedia.org/wiki/Nothing-up-my-sleeve_number , e.g. digits of pi, 12345..., or the output of an unrelated hash function) is revealed. If no such preimage is revealed by the end of 2027, even if practical attacks are theorized or in progress, this market will resolve as "NO".