We all know she's broken. Whether you think so or not, it's stupid that all four of her abilities can kill at all, some of them being very likely to do so. So I've thought up some potential reworks for everything, assuming Riot is still hell-bent on keeping her in the game.
Her boom bot is roughly balanced I would say. I acknowledge that it meets a standard for being "utility," being that it serves a purpose as a room or alley clearer. The only difference I would make would be to decrease the radius of the explosion. Even if it's just a little bit, I think it's a warranted fix. You should be punished for sitting on your rear end while that thing is running towards you, but if you actively make an attempt to get out, it shouldn't tag you unless you started to run too late. A slight radius decrease for the boom bot explosion and that's it. Perhaps more health on it, too, but that's up for debate.
The blast packs have, like the boom bot, an apparent use for utility purposes. The ability to boost oneself up and forward is quite useful and, I think, balanced. But that should be as far it gets. It makes sense that an explosion would do at least some damage, but 75 is too much. Two hits and a full health opponent is down. Even if you don't get full damage, it's still too good because it takes half a second to release at almost no cost to the Raze mid fight. I suggest lowering the damage to 20 or 30. I know it's a big difference, but I think it focuses that ability far more on utility and it's still useful in some cases for picking off already weakened players at the end of a round right around a corner.
As for the cluster grenades, I would argue that they aren't even utility. Sure, in CS:GO "utility" included grenades, but I think "utility" in Valorant should be seen in a different light. If we take any of the other characters, each and every single one of their abilities serves a strategic purpose. Let's run through a scenario. As Brimstone (or Phoenix or Viper), I want to ensure that my opponents can't stand on or push through location X. So I pop an ability (molly, fireball, or poison). That will keep them at bay for a good deal of time and it can be used strategically. They would have to suffer from mental vacancy to remain on my ability, or my teammates would have done their part in forcing them into uncomfortable positioning. Either way. it's not a guaranteed kill, nor does it do significant damage. It has a strategic purpose. Whereas Raze's grenades are quite the opposite. They're basically an insta-kill if you catch someone with them and if it's not an insta kill, then it gets unnervingly close. There's no strategic use to them. One might argue that she could use them to clear a room. I would respond, saying that it's not fair that something so trivial as one ability can clear an entire room or at least significantly damage most areas of the room. The radius is a huge issue. You could be so seemingly far from the initial or even the secondary explosion and it would still tag you for an obscene amount. Making the ability one charge instead of two (though there is still the potential of gaining a second) didn't really solve the issue. The issue was with the mechanics of the ability itself, not with the fact that there were two of them, though two at once were honestly quite annoying. I suggest decreasing the radius of the explosions and also adding a bit more fuse time. If the fuse was just slightly longer, it would make a seriously beneficial difference.
As for her ultimate, it needs one of two things. It needs either to be 8 ult charges or it needs to do some recoil damage to Raze herself. Perhaps just 20 or 30, but for an ult with that much potential and effectiveness (even mid fight), it needs to have greater cost. High risk high reward, or just make the reward harder to obtain.
These are my suggestions, feel free to add your own and debate them. Another potential suggestion of mine is to rework her cluster grenade and ult entirely, but I assume, based on her character traits, that wouldn't be an option.
Good day to all.