How many people play GTA RP? The honest answer is layered

9 min readUpdated 8 October 2026

gta-rproleplaynopixelfivemdatatwitchkick

Correction, 8 October 2026. This post was built on the player count published on FiveM’s front page, and that counter is not live. Re-checked on 8 October, the page returned the identical figure — 120,520 online now — four days after we first read it, across several cache-busted requests, and it is served from Akamai NetStorage with a last-modified date of 2 September 2026. The number is baked into a static page, so it is a snapshot of some unknown earlier moment rather than a reading of the moment you load it. A Cfx-operated endpoint, runtime.fivem.net/counts.json, returned 297,210 at 04:09 UTC on 8 October; we cannot confirm it measures the same population, so we are not substituting it as the answer. The headline and description have been changed accordingly. Everything below that comes from our own census is unaffected, and the sentence in “What this does not show” that flagged the counter’s unknown refresh interval was right to be suspicious — it just did not go far enough.

120,520.

That is what the front page of fivem.net said at 04:16:14 UTC on 4 October 2026 — “Join the 120.5k people playing right now”, with the exact figure sitting in the page markup as 120,520 online now. It is the closest thing to an official count that exists, because Cfx.re, which runs FiveM, has been part of Rockstar since August 2023.

Two things immediately spoil it as an answer. It counts everyone on any FiveM server — roleplay, racing, drift, deathmatch, an empty box someone left running — not everyone playing roleplay. And it excludes the roleplay server that most of the world is currently watching, which runs through Rockstar’s own launcher and needs no FiveM client at all.

So the useful version of the answer is layered, and the second layer is the interesting one.

The number Google gives you is from 2021

Search this question and page one leads with a PCGamesN piece reporting that FiveM “hit a peak of 250,000 concurrent players over the weekend.” It was published on 13 October 2021 and the tweet it cites is from April that year. Five years on it is still the top measured number anyone offers.

Nobody publishes a current global figure because there is nothing to publish it from. Roleplay is split across hundreds of independently run FiveM servers, Rockstar’s launcher-native nopixel V, and standalone communities with their own clients and their own counters. No one aggregates them, and the operators who could have no reason to.

What you can read today, from the counters that do exist and publish:

counter reading, 04:16 UTC 4 Oct 2026 what it actually measures
FiveM (Cfx.re) 120,520 anyone connected to any FiveM server
GTA V Enhanced on Steam 43,129 sessions launched through Steam
GTA V Legacy on Steam 29,613 sessions launched through Steam

The Steam figures come from Valve’s own GetNumberOfCurrentPlayers endpoint at 04:16:31 UTC. Do not add these up. Steam counts sessions it launched; a FiveM session is launched by the FiveM client, and we have no way to measure how much the two overlap. What the table does support is the comparison that has been true on and off since 2023: at that minute, FiveM alone had 1.66 times as many people connected as both Steam editions of GTA V put together.

One caveat we would want applied to us, so we will apply it here: 04:16 UTC is a quiet hour. Across the thirty days to 04:00 UTC today, the mean concurrent GTA viewership we measure at 04:00 is 239,949 against 506,545 at 21:00, the daily peak. This is not FiveM’s high-water mark and should not be quoted as one.

What we can count precisely: who is watching

We sample every live Grand Theft Auto stream on Kick, Twitch and YouTube once a minute. At 04:15:43 UTC on 4 October 2026 there were 171,032 concurrent viewers across 1,948 live channels, and roleplay was the majority of it.

viewers share channels share median channel
Roleplay 95,544 55.9% 838 43.0% 7
Everything else 75,488 44.1% 1,110 57.0% 5
All live GTA 171,032 100% 1,948 100% 5

That majority is new. On 25 August, at the same hour, we measured roleplay at 36,944 of 198,679 viewers — 18.6% — on 812 of 2,054 channels. Six weeks later the roleplay audience has nearly tripled while the category as a whole has fallen about 14%, and the number of roleplay channels has barely moved: 812 to 838. None of the growth is more people streaming roleplay. All of it is more people watching the ones who already were. The mean roleplay channel went from 45.5 viewers to 114.0.

Most of that audience is watching a server you cannot join

Eighty-one of those 1,948 channels had “nopixel” in the title. Between them they held 65,708 viewers — 38.4% of everyone watching GTA anywhere, and 68.8% of everyone watching roleplay.

group viewers share of all channels median channel
nopixel V 65,708 38.4% 81 76
All other roleplay 30,516 17.8% 758 6
Everything else 74,808 43.7% 1,109 5
Viewers — 171,032 concurrent 38.4% 17.8% 43.7% Channels — 1,948 live 4.2% 38.9% 56.9% nopixel V other roleplay everything else

The median nopixel channel had 76 concurrent viewers. The median channel streaming any other kind of roleplay had 6. Same game, same hour, same category pages.

And here is the part that makes the opening question hard to answer honestly. nopixel V entered closed beta on 8 September, running through the Rockstar Games Launcher rather than FiveM, with access limited to roughly 450 invited creators and established roleplayers — reported by GTA BOOM and others, not something we can verify ourselves. There is no public application form, and as of this week no open-access date has been announced.

About 120,000 people can log into FiveM roleplay right now. A few hundred people can log into the roleplay that 38% of the audience is watching.

The servers you can apply to are all there in the same census, and they are small by comparison: Prodigy on 61 channels and 3,508 viewers, The Academy on 11 channels and 683, ONX on 16 and 468, Bloodlines on 5 and 304. Every public roleplay server we can identify by name, added together, is a rounding error against one invited beta.

Of the 81 nopixel channels, 52 were on Twitch (45,568 viewers) and 29 on Kick (20,140). Eighty of the 81 carried an English language tag, which is worth noticing on a platform mix where Arabic-language creators take a far larger share of total GTA viewing than their channel count suggests.

What this does not show

We count viewers, not players, and the two are not the same population or even the same order of magnitude. Nothing in our data tells you how many people are in a roleplay city — only how many are watching someone else be there.

Server identification is a title keyword match, so every server figure above is a floor. A nopixel streamer who titles their stream with only their character’s name is counted as generic roleplay, and the same goes for Prodigy and the rest. Our content-type classifier reads titles too, with freeroam as the fallback bucket; a hand-check on 4 August found 30 channels and 7,081 viewers whose titles named a roleplay city and which the classifier had filed as freeroam. Read 55.9% as “at least”.

This is one hour of one day. Four complete censuses between 04:08:41 and 04:15:43 UTC put roleplay’s share between 55.7% and 56.9%, so the reading is stable across eight minutes — not across the clock, and 04:15 UTC misses Asian primetime entirely. Today’s 04:00 hour is itself running below its own 30-day mean, so if anything these are quiet-hour numbers.

Discovery depth differs by platform: we page much deeper into Twitch’s directory than Kick’s, so channel counts are not comparable across platforms and the medians above should be read within a platform, not between them. Viewer totals are not affected the same way.

And FiveM’s counter is theirs, not ours. We do not know how often it refreshes — it read the same 120,520 at 04:16:14 and again at 04:17:30. It read the same 120,520 again on 8 October, which is what prompted the correction at the top of this page: the figure is published as static markup, not as a live reading, so it should not be quoted as a current concurrency.

Why any of this matters on 19 November

GTA VI ships on 19 November 2026 for PlayStation 5 and Xbox Series X|S. That is what Rockstar’s own page lists, and it listed nothing else when we checked it this morning. There is no announced PC version.

Roleplay is a PC phenomenon end to end. FiveM is a PC client. nopixel V runs through the Rockstar Games Launcher on PC. So 55.9% of live GTA viewing right now is in a form that GTA VI does not ship with on day one, and the majority of that is concentrated on a single invited server. Whatever happens to the category in November, the roleplay half of it is not moving — not until there is a PC version and a modding framework on top of it, and neither has a date.

Cite this

Every viewing figure above comes from a single complete census of every live GTA stream on Kick, Twitch and YouTube at 04:15:43 UTC on 4 October 2026, with the stability check drawn from three further censuses at 04:08:41, 04:11:41 and 04:14:41. In all four, the top-400 board and the tail file carried matching timestamps, the per-stream counts reconciled exactly to the published totals, no channel appeared twice, and no platform reported degraded coverage. The hour-of-day means are the 721 hourly points in live/summary.json covering 04:00 UTC 4 September to 04:00 UTC 4 October. The FiveM figure was read from fivem.net at 04:16:14 UTC and the Steam figures from Valve’s public player count endpoint at 04:16:31 UTC, both on 4 October 2026 — but see the correction at the top of this page before quoting the FiveM figure, which is published as static markup and is not a live reading.

Permalink: https://luciastream.com/blog/how-many-people-play-gta-rp/. The sampling minute matters as much as the number here — roleplay’s share moves with the clock, so quote both. How we collect it is written up, the live board is public, and the thirty-day creator leaderboard is where the nopixel channels above can be checked against their own history.


Numbers in this post come from our own sampling — seehow we measure.