The entry level SDG2000X model is almost exactly twice the price of the DG811 (which is -- assuming the hack is applied -- performance-wise virtually identical with the DG2000). And the SDG2000X, albeit it's a great performer, isn't without its flaws either.
IMO, currently the DG811 ("improved") is the sweet spot value-for-money wise for a "serious" entry level AWG that works as supposed right out-of-the-box. The cheaper "Feeltech" or similar AWGs require some modifications to achieve similar level of performance and still may never match the professionally built units, let alone mechanics-wise (solidity of casing, EMI shielding, cooling approach etc.). As a bottom-of-the-barrel unit -- if the hobby budget doesn't allow anything else -- they may get the job done, though.
The next (small) step up from the DG800 may be the Siglent SDG2000X that's got a reputation of providing exceptionally clean output signals (distortion-wise), but on the other hand shows its age by the rather old-fashioned appearance of its user interface and the half-hearted utilization of the touch screen. The DG800 also provides a decimal input keypad, though it's accessible via its very well implemented touch screen instead of a tiny, physical silicone membrane keypad. Spec- and performance-wise, the fully "liberated" DG800 (~DG992) and the SDG2042X (SDG2122X) don't differ that much, the maximum frequency (sine) is 100MHz vs. 120MHz, and the "real" sampling rate is 250M vs. 300M (don't get confused by the maximum "oversampling" rate of 1.2G thrown into the ring by Siglent's marketing department, this has not much significance in the real world). The metal enclosure of Siglent's AWG is much more attractive, though, but since it's got a fan installed, its accoustical footprint has some effect in the lab while Rigol's entry level AWGs are completely silent.
There's one important drawback of both Rigol's and Siglent's AWGs: The output ground is earth-referenced and thus requires an isolation transformer or -amplifier (depending on the signal's parameters to be fed to the DUT) in order to break a possible ground loop.
In the end, everyone who's in the market for a new entry level AWG should ask his "crystal ball" what the important spec figures of his new goint-to-be instrument are, check the available offers and maybe his personal preference, decide wether hacking is an option or not (no big deal in both mentioned cases, can be applied by external means and it's fully reversible) and then place the order...