Between the two, I only have the Lumintop HL3A.
I chose it mainly because I love the stepless ramping of the FW3A.
The HL3A is a great headlamp, but there are others that might be better suited to you.
Wow I thought the HL3A was light, but the H04 is almost 15% lighter (56 vs 48g). Factoring in the weight of a cell, 106g vs 98g makes it an 8% difference. The inclusion of magnetic charging is pretty convenient and that efficient buck driver makes runtimes excellent. Brightness suffers though compared to the HL3A, Zeroair tested the max brightness of the H04 with the LH351D to be just 600 lumens while the max brightness of the SST20s in the HL3A with full fet should be 1600+.
Note that I am biased towards the HL3A which I own.
H04 RC Pros:
Better efficiency- maybe 10-20% more runtime at equal brightness (my estimate)
Magnetic charging
Lighter
Better headband
HL3A Pros:
More feature rich UI (anduril)
Higher turbo brightness
Better tint
Better colour rendering 95/80 of sst20s(Reds really pop) vs 9050 of the LH351Ds
I estimate that both headlamps should be able to a similar consistent brightness of 350 or so lumens in their high cri configurations. While the H04 has a more efficient driver, the HL3A does have the advantage of having 3 LEDs.
I like the ramping and features of anduril too much and the sst-20s make everything look gorgeous. I’d pick the HL3A again, and for me the H04 would be the next best choice.
If you care about efficiency, Skilhunt H04 all the way as it uses an efficient buck driver. In higher modes, this may be offset by the HL3A being a triple.
The HL3A may have FET for a higher turbo, but that will overheat and step down in seconds, so I wouldn’t let that influence your decision. Fun feature to have on a handheld flashlight where you have your hand next to the switch, but almost useless in a headlamp that is intended to be hands free.
Thank you for answers, I have 2x Sofirn IF25A to have fun with Anduril so now I will go with something other 
Thanks for every help, haven’t order yet so you can still write if you have something to add about one of these, I still can change my mind
I got the H04R (OP reflector) and then ordered a TIR via email from skilhunt. It was nice to be able to try both. I like the TIR a bit better since the beam is more usable overall, but the OP reflector is alright for outside use. You could buy the OP one, and I can send you an extra TIR that I have (if you are in US pm me).
I’ve got few Skilhunt headlamps and I love them, highly recomend. I prefer the TIR versions. Spare TIRs are so cheap that you can order few of them (varying degrees) and see wich one suits you.
@NoDa775
Thanks for your offer but I’m from Poland so the only way for me is buying from Aliexpress.
@Mraz
I think I will go for TIR.
Is it possible to swap between OP and TIR or there is bigger difference than only this plastic in front.
If you want the most throw, the OP is the way to go. If you buy the OP, you can easily find TIR optics on Ali Express and swap them in. If you buy the TIR version, it will be very difficult to source the reflector if you decide to go that route.
I am very happy with H03RC. Indeed that was my first LED flashlight which ignited me in that hobby. I want to buy H04 but little different model H04F, I really don’t need magnetic charger. With previous F models I know there was a issues. If someone have that new F models with additional TIR cover please comment. And now I am happy just ordered H04F Hi cree for just 30$ from AE 
Then def go with the TIR option, much smoother beam overall and you can find TIR with narrower beam angle as a backup.
Yes with 3$ AE coupon. But without it there is promotion for about 35$.
It was code only for you or everyone can use? Could you give me if it’s for everyone?
The coupon I think was exchanged for AE coins but I am not very sure about that. I didn’t remember from where I received it.
My two cents as a longtime Zebralight H60 user ,and Sofirn SP40 user of about a year. I expected the UI of the Skilhunt H03 SE (recently purchased) to be better given that I had heard nothing bad said about it. I think it would be much better for my brain if the mode order was “cyclical” rather than “reciprocating”. I end up having to make an awful lot of clicks in the process of changing brightness. If you don’t need to change brightness often the UI will do the job. On the plus side it is very light (good job Skilhunt), as light as a Zebralight. I wanted the LH351D version but didn’t want to pay the substantially higher price found on Skilhunt’s store versus an XPL 5000k version on Banggood with a coupon.
I don’t like cool white low CRI (it’s noticeably cooler than my Zebralight neutral white) but it does well as a daytime headlamp and is still a lot better than a budget headlamp.
I am sure Anduril would work very well for headlamps, I have been tempted to add a HL3A or a PL47G2 to my collection.
The YLP Gekko 1.0 is the bee’s bollocks in this price range. 4000K LH351D 90+CRI, ramping or discrete/good programmable UI, very good thermal control, 49 g, quite a bit more efficient than the H04. I was cross researching against the HL3A, H04/H04RC, YLP Panda 2 CRI/3, some Acebeams, etc. The HL3A was initially the most attractive, but button design and Lumintop design/quality issues knocked it down. I decided to jump the gun before the reviews rolled in to get the initial discount price from the code in the YLP thread here (with code was $30 USD, $36 USD shipped to USA, currently ~$41 USD not shipped without coupon code if you do the conversion from rubles), and I’m very satisfied.
There is some criticism of the heat path/build quality between the star and body, but, if you don’t open it up to see, the construction looks impeccable externally on mine. It’s not super high powered, 3 amps max, so the heat path issue actually doesn’t affect use to a noticeable degree. Mine actually does sustain 900 lumens in actual, cooled (running) outdoor night use on a cool night. 300+ lumens is sustainable uncooled indoors.
AEDe has a nice review on Fonarevka.
Nice 300 lumen flat line runtime for 4.5 hours using a 3500 mAh cell:
http://forum.fonarevka.ru/showpost.php?p=1331202&postcount=173
Much better efficiency than the H04RC (lumens*hr, area under the graph) at 350 lumens from zeroair’s review using a 3400 mAh cell. Both headlamps use the LH351D, though the Gekko is 4000K and the H04 is 5000K. If everything else was equal, you’d expect the 5000K to be more efficient than the 4000K, so the efficiency gap between the drivers is pretty big.
H04 have same hi-cri 4000K led, superior build quality and switching driver which is more efficient. You can compare Gekko to this one, they probably are made by same manufacturer because have same hole under star
It is not correct to compare efficiency on 300 lumes and 350 lumens, moreover on the 300lumens real output could be 280.
But stabilization is really flat.
Bt you shold take in accont quirks in Gekko TK.
As example, if you turn Max (900) and work for some time (TK will lower output as example 400lumens), and switch for 300lmens mode, the real output would be significantly lower (~180-200) for 1.5 mins while TK will raise it up to 300 lumens. Because it’s “TK overheat gap” d’t take in account that 300lumens mode by default is lower heated then previos “downlowered maximum”.
Other - it would be very good to check how good LED star attached to body. There was case of overheating LED and not good screwing.
Not correct, because of driver difference. And head heat dissipation differs, but in other body is pretty close. Interesting case.