FiveM Enhanced stress tests start 14 August

by 6Charts Team Category: news 7 min read

We read the Cfx Forum announcement, the July development update and two GitHub patch note threads at source on 13 August 2026. The first test runs 14 August from 17:00 to 19:00 UTC, with player slots increased gradually through the session. Cfx.re never names a target player count in the announcement, so the link to the 2,048 cap is our inference and we label it as one.

Cfx.re opens a run of public stress tests for FiveM for GTAV Enhanced tomorrow, 14 August, from 1 to 3 p.m. ET, which is 17:00 to 19:00 UTC. Anyone running a server, or planning to, has about a day to decide whether to take part. The short version for operators is that this is load testing with volunteers, the slot count climbs during the session rather than starting at the top, and the announcement itself is quieter about numbers than the coverage around it. What Cfx.re actually announced CONFIRMED (our own retrieval, August 13, 2026): we read the announcement on the Cfx Forum, topic 5420662 in the Announcements category, posted by the staff "system" account at 2026-08-12T13:58:22.962Z. It showed 11,886 views and 22 posts when we pulled it. The post opens: "We will be hosting a series of public stress tests for FiveM for GTAV Enhanced. You can join these tests on the following dates and times: August 14 from 1-3 p.m. ET (17:00-19:00 UTC) ... September 4 from 1-3 p.m. ET (17:00-19:00 UTC)". Our capture of the date list preserved the first and last entries and elided the ones between, so we can confirm the series starts on 14 August and the last listed date is 4 September, and we are not going to guess how many sessions sit in the middle. On what the sessions are for, verbatim: "Each test will run for approximately two hours, and we will be asking players to help by performing specific actions during these tests. These tests will enable us to test certain behaviors and server performance at scale. Player slot capacity will be gradually increased as we progress through the session." Two practical details from the same post. Cfx.re says it "also created a read-only Discord channel called #public-stress-tests", which is where instructions and, we assume, the connect address will appear. And there is a reward: "Stress test participants will also be awarded a unique Cfx Forum badge for their efforts: this will be granted automatically after each session." Where the 2,048 figure comes from The 12 August announcement never says 2,048. It never names any player count at all. The number comes from a different post. CONFIRMED (our own retrieval, August 13, 2026): Cfx.re Development Update #3, forum topic 5415045, posted 2026-07-15T14:59:37.412Z, 35,333 views. Verbatim: "With these foundations in place, we're now looking to push beyond the current 2,048-player cap on a single server. We'll be scheduling public playtesting sessions in the coming weeks to test at higher player counts. Stay tuned for how to participate." So July promises public playtesting sessions to test above 2,048, and August delivers public stress tests. Tying the two together is our inference, and we want to be clear that it is one. Cfx.re has not confirmed a target slot number for 14 August, has not said the cap will be exceeded during this series, and the only capacity language in the announcement is that slots will be "gradually increased" through each session. If you see a headline promising a specific player count tomorrow, that number did not come from Cfx.re. What changed under the hood Development Update #3 is the substantive document, and its technical claims are worth reading before deciding whether the tests matter to you. All of the following are CONFIRMED (our own retrieval, August 13, 2026) from that post. Server update rate: "Server operators can now configure it up to 120 updates per second, up from the previous 30 (or 40 with sv_useAccurateSends enabled) on FiveM for GTAV Legacy." Transport: "Sync data is now transmitted as raw UDP packets (with ENet still handling event-based communication)", which Cfx.re says opens the door to "Registered I/O (RIO) on Windows and io_uring on Linux." Culling, described as it used to work: "Previously, it ran on the sync thread and could only process around 16 players per tick. On a server with thousands of players at 30 Hz, it could take several seconds to cycle through everyone." Memory: "memory allocation per player and per entity has been optimized (reducing RAM usage by up to 50%)". The culling line explains why high population servers on Legacy feel the way they do. If it takes seconds to walk the player list, some players are being told about other players late, which is the mechanism behind a lot of pop-in complaints at scale. A number correction from the thread CONFIRMED (our own retrieval, August 13, 2026): the reaction in the announcement thread is mixed. BELIALS_CUSTOMS wrote at 2026-08-12T22:32:38.830Z: "GTA V Enhanced is the same as GTA V Legacy-better in some things, worse in others. Graphically, I barely notice the difference, so I don't know what 'craziness' we're talking about other than giving headaches to modders like me, making us migrate hundreds and thousands of hours of work just so it looks a 'little bit' better." AJTheDev replied at 2026-08-13T03:32:30.697Z: "Yeah because a doubled tickrate and better networking is just looking a 'little bit' better?" The substance of that reply matches what Cfx.re published, but the multiplier is low. By Cfx.re's own figures the ceiling moves from 30 updates per second, or 40 with sv_useAccurateSends, to a configurable 120. That is up to four times, or three times against the accurate sends baseline. It is a bigger change than the word "doubled" suggests, and operators budgeting CPU headroom should plan against the larger number. An official Grand Theft Auto VI screenshot. The stress tests are for FiveM on Grand Theft Auto V Enhanced, and Cfx.re has not connected them to GTA VI. One request went unanswered. PhilBB wrote at 2026-08-12T23:42:09.825Z: "Sign me up...I am looking forward to escrow being included in this test please CFX !!" Edxtti_x followed at 2026-08-13T03:25:03.696Z with "bring over escrow". No Cfx.re staff member replied to either. The only staff activity in the thread beyond the original post was an image from the account "cobalt" at 2026-08-12T15:05:49Z. Escrow support is therefore not in scope as far as the public record goes, and we are not treating silence as a no. What operators should check before tomorrow CONFIRMED (our own retrieval, August 13, 2026): two early access patch note threads on GitHub matter more to server owners than the stress test itself. citizenfx/rfc Discussion #395, "FiveM for GTAV Enhanced: Early Access Patch Notes August 11", posted 11 August 2026 by Cfx Admin, includes the line "Diagnostics and metrics endpoints now require authentication". That is the most disruptive item on the list. If you scrape those endpoints into Grafana, Prometheus or any other monitoring stack, expect that pipeline to go quiet without an obvious error. The same thread adds "generate_handshake_key and load_handshake_key commands, so the handshake key can be generated and loaded from a file", "a metric for execution time per sync stage", and a fix so that "Setting a player's routing bucket to a very high id with SetPlayerRoutingBucket no longer makes the server reserve all system memory (#391)". citizenfx/rfc Discussion #404, "FiveM for GTAV Enhanced: Early Access Hotfix 9", posted 12 August 2026 by Cfx Admin, adds a SetPlayerName server native with an accompanying onPlayerNameChanged event, new onPedDeath and onPedHealthChanged events, integer coordinate support in CreateColshapePolygon, and a fix so NUI loads non-.woff2 fonts such as .ttf and .otf again. A commenter in that thread asked for the parameters on the new ped events, which is a fair complaint: they shipped without documentation, so anyone adopting them is reading signatures by trial. What we could not reach We did not join the read-only Discord channel, so we do not have the connect address and cannot tell you how to get in beyond following that channel. We could not establish whether tomorrow's session actually exceeds 2,048 slots, because the announcement commits only to gradual increases. We found no documentation page for the new metrics authentication, so we cannot say whether it is opt-out, what the auth scheme is, or whether existing setups break by default. And we did not confirm build numbers for the 11 August patch; b115 and b116 circulate in third party bug reports and we treat them as unconfirmed. What we confirmed and what we did not CONFIRMED (our own retrieval, August 13, 2026): the 14 August start date, the 17:00 to 19:00 UTC window, the 4 September end of the listed series, the gradual slot increase, the read-only #public-stress-tests channel and the participation badge, all from Cfx Forum topic 5420662 posted 2026-08-12T13:58:22.962Z. CONFIRMED (our own retrieval, August 13, 2026): the 2,048-player cap language, the 120 updates per second ceiling, the raw UDP sync change, the 16 players per tick culling description and the up to 50% memory reduction, all from Cfx Forum topic 5415045 posted 2026-07-15T14:59:37.412Z. CONFIRMED (our own retrieval, August 13, 2026): the metrics authentication requirement, the handshake key commands, the sync stage metric and the routing bucket fix from citizenfx/rfc Discussion #395, and the SetPlayerName native and ped events from Discussion #404. OUR INFERENCE, not a Cfx.re statement: that these stress tests are the sessions promised in July for testing above 2,048 players. The 12 August announcement names no player count. OUR CORRECTION: the tick rate change is up to four times the Legacy default of 30, or three times the 40 available with sv_useAccurateSends, rather than doubled. UNVERIFIED: the connect address, whether tomorrow exceeds 2,048 slots, whether escrow is in scope, the metrics auth scheme and whether it breaks existing monitoring by default, and the build numbers b115 and b116. WE DECLINE TO CLAIM: that Cfx.re has committed to any specific player ceiling for this series, or that staff silence on escrow means escrow was rejected.