It's very unlikely anyone was doing any research on directly using SARS-CoV2 as a weapon. It kills or maims too low a percentage of people to have tactical value, and it's too difficult to contain. (The most effective weapons severely handicap their victims and allow them to live into old age, taking fighters off the field, and turning them into long-term liabilities and living reminders for anyone who might think about fighting you in the future.)
I'm not saying SARS-CoV2 leaked from a lab, but if it did, it was probably more of a basic science/weapons background research rather than an engineered weapon itself. You might want to add some SARS-CoV2 characteristics to a bioweapon, but you'd want to start out with something with greater morbidity and more easily quarantined as a starting point for a weapon.
I do not believe that covid was intentionally designed and released as a bio weapon.
That being said, have you heard of long covid?
Why should I believe this is any more real than "chronic lyme"? There are a whole lot of hypochondriacs out there; something proponents of "long covid" and "chronic lyme" never seem willing to acknowledge.
The groups promoting both of these organize and operate the same way, and make similar claims. Huge lists of nonspecific generic symptoms and facebook groups full of uncritical believers mutually reinforcing each others' beliefs (parallel to the well understood phenomena of "support groups" which promote eating disorders and create social feedback loops for reinforcing/worsening body dismorphia.)
History says you are wrong to discount NATO countries (I include Japan as an unofficial member) using bio-weapons. They have a long history of deploying and supporting deployments of these kinds of weapons against military and economic foes.
You seem to assume a bio-weapon has to cause mass death to be effective and meet the deployer's objectives...you are wrong in the case of economic attacks.
The number of people with long COVID symptoms is a tiny tiny fraction of those exposed to SARS-CoV2. If it's a designed feature of SARS-CoV2, it's very poorly implemented, unless it's actually very specifically targeting some as-of-yet unidentified demographic. (This seems very unlikely.)