After testing and analyzing multiple firmware versions on the GL-MT6000 (Flint 2), I wanted to share some constructive feedback and expectations regarding the upcoming v4.9 firmware release.
This post is based on:
real-world testing,
firmware analysis,
community discussions,
comparisons between:
v4.8.3 stable,
v4.8.4,
v4.9 beta,
and OpenWrt builds.
First: Thank You for Continuing Development
The MT6000 hardware is excellent:
MediaTek Filogic platform,
strong CPU,
excellent 5 GHz capabilities,
very good thermal design,
and overall one of the best OpenWrt-capable routers in this class.
The upcoming v4.9 firmware also looks promising because it introduces:
QoS/SQM integration,
DPI/security visibility,
IoT network support,
improved UI modernization,
better traffic management,
and general optimization features.
According to community discussions and preview information, v4.9 focuses heavily on:
The Main Concern: Please Avoid What Happened in v4.8.4
Many users noticed that firmware v4.8.4 felt significantly more restrictive regarding Wi-Fi behavior compared with v4.8.3 and OpenWrt builds.
Common reports included:
weaker 5 GHz coverage,
reduced signal penetration,
lower throughput at distance,
less stable repeater/mesh links,
and overall more conservative Wi-Fi behavior.
There have also been several community discussions about 5 GHz performance regressions and users preferring older firmware versions because of better Wi-Fi behavior. oai_citation:1‡GL.iNet Official Forum
Firmware Analysis Findings
After comparing:
v4.8.3 stable,
v4.8.4,
v4.9 beta,
and OpenWrt builds,
it appears that:
v4.8.4 introduced:
more aggressive regulatory enforcement,
more conservative DFS behavior,
tighter wireless tuning,
and possibly stricter thermal/radio protection behavior.
Meanwhile:
OpenWrt builds appear more liberal regarding TX behavior,
while v4.9 beta appears more optimized and balanced than v4.8.4.
Important Observation About TX Power
On official firmware:
many users observe:
Tx-Power reporting around 20 dBm,
even on:
UK region,
DFS channels,
Channel 100,
HE80.
Meanwhile:
OpenWrt builds on the same MT6000 hardware have reported much higher TX limits.
This strongly suggests:
the hardware itself is capable,
but firmware/regulatory handling is more conservative on official releases.
Why This Matters
The MT6000 is often used for:
large homes,
mesh/repeater setups,
long-distance 5 GHz coverage,
VPN gateways,
SQM/QoS environments,
and advanced networking.
Users buying a flagship router expect:
strong 5 GHz performance,
excellent coverage,
and stable high-throughput wireless links.
What Many Users Hope v4.9 Will Deliver
The current v4.9 beta already looks more promising than v4.8.4 because it appears:
better optimized,
smoother,
more balanced,
and less aggressively restrictive.
Many users would love to see v4.9 become the best balance between:
I would be curious to hear what the reason for the wireless changes with the newer firmware (because its my understanding this impacted multiple models). If it was because of the FCC requirements and they were shaping up their firmware to be compliant because of this:
Then it begs the question why werent they following a countries standards in the first place when they released the router to the countries(s) in question
That is actually a very good question, and after testing multiple firmware versions on the GL-MT6000, I also became curious whether the newer wireless behavior is related to regulatory compliance changes.
From what I observed:
v4.8.3 behaved much less restrictive,
v4.8.4 became significantly more conservative,
and v4.9 beta appears more optimized again, although still capped around 20 dBm reporting.
What makes this interesting is that official firmware stays limited even when using:
US region,
Channel 149,
channels where FCC regulatory database itself allows much higher limits.
Meanwhile OpenWrt builds on the same hardware were able to report much higher TX values.
If these changes were indeed related to FCC/CE compliance concerns, I think many users would simply appreciate transparency from GL.iNet explaining:
what exactly changed,
why it changed,
and whether the older firmware behavior was considered non-compliant or simply less conservative.
Because from a user perspective, the difference between 4.8.3 and 4.8.4 was noticeable enough that many people downgraded firmware specifically because of Wi-Fi performance and coverage concerns.
I think most advanced users understand the need to comply with regulations — the main request is simply:
clear explanation,
consistency,
and the best possible wireless performance within legal limits.
I am not sure what is reported in the GUI is actually what is happening for the US (I know there is clearly a big change in EU or at least DE):
I have used 4.8.3, 4.8.4 and 4.9 beta on my Flint 2 and am located in the US and have the US set as region. I have also tried Openwrt 25.12.
I am using channel 153/80 on 5Ghz
While Openwrt 25.12 reports 30dbm and the Gl.Inet Firmware versions do not I have not seen any difference in actual dbm levels on the receiving end (meaning the mobile device) when further from the router between any of these versions.
I have seen worse performance on OpenWrt 25.12 - I suspect the beamforming algo in the pro drivers is better optimized to the actual antenna geometry than the generic on
While Openwrt 25.12 shows 30dbm and Gl.Inet shows 20dbm it looks to me that the actual transmit power is the same.
The same observation is also true for me on 2.4Ghz - I have some devices that are quite far from the router and they report the exactly same signal levels on any 4.8.3, 4.8.4 and 4.9
Again - this observation applies to a US bought router with US set as the region and not to Europe.
Have you actually measured performance differences and seen lower signal levels on the mobile device in exactly the same location on the same channel?
That is actually a very valuable observation, and I think you may be correct that the reported TX value itself is not telling the full story, especially in the US region.
My own testing also started to suggest that:
the official GL.iNet firmware may be reporting a conservative conducted TX power value,
while the effective real-world wireless performance may still be similar because of:
beamforming,
antenna tuning,
EIRP handling,
and MediaTek proprietary driver optimizations.
I also agree with your point regarding OpenWrt:
despite reporting 30 dBm,
actual real-world coverage did not always feel stronger,
and in some cases stability/performance was actually worse.
This is why I started suspecting that:
GL.iNet/MediaTek proprietary drivers may simply be better optimized for:
the MT6000 antenna geometry,
beamforming behavior,
chain balancing,
and thermal/radio management.
So the 20 dBm shown in iwinfo may not represent the actual effective RF behavior users experience in practice.
Where I still think there is something interesting though is:
many users (especially in EU/DE regions) reported noticeable differences between firmware versions,
particularly around 4.8.4,
with weaker long-range behavior and lower perceived coverage.
So there may still have been:
regulatory tuning changes,
DFS behavior changes,
or more conservative wireless algorithms introduced in that release.
I have not yet done a fully controlled scientific RSSI comparison with:
same client,
same exact location,
same channel,
across all firmware versions.
So at this point I would say:
there is strong evidence that reported TX power alone is misleading,
and real-world wireless performance depends much more on driver tuning and beamforming optimization.
Your observation about OpenWrt vs official firmware performance is extremely useful because it supports that theory quite well.