Skip to content
LazyEloGet the app
Menu

CS per Minute in League of Legends: Benchmarks and How to Improve

Updated October 29, 2026 · 9 min read · by the LazyElo team

CS, creep score, is the number of minions and monsters you kill. It is your most reliable income: kills come and go, but every wave brings gold and experience whether the game is going well or not. This guide covers what a good number looks like for your role and rank, what the popular targets get wrong, and how to actually raise yours.

CS per minute or CS at 10 minutes?

CS per minute is your total CS divided by the length of the game. It is the number on most post-game screens, and it is fine for comparing whole games. But it mixes two very different skills: last-hitting in lane, and catching side waves later on. A long game where you spent fifteen minutes grouping around baron drags it down even if your laning was perfect.

CS at 10 minutes isolates the laning phase: same map, same opponent, mostly the same waves for everyone. It is the better number to train, and the one this guide focuses on for laners. Junglers count camps rather than waves, and supports aren't meant to farm at all, so both read the number differently (more below).

Why farming decides games

In an analysis of ranked games at the ten-minute mark, the gold and experience differences between the teams stood far above every other variable as predictors of the winner. CS is the biggest part of both. A kill is worth more than one minion, but CS is the part of your gold you can earn every 30 seconds without taking a risk.

You can see it in a single player's games too. In the ranked lane games of LazyElo's developer (415 games in top, mid and bot), being 10 or more CS ahead of the lane opponent at 10 minutes went with a 57% win rate (138 games), and being 10 or more behind with 41% (83 games). One player is not a law of nature, but the gap is the same direction as the large studies, and it is a gap you build one last hit at a time.

Benchmarks per role and rank

Here is the surprising part: CS at 10 doesn't change much between ranks. League of Graphs data for a typical top laner shows CS at 10 rising by only about 4 to 9 CS from all ranked players (“Iron+”) to Master and above. For Jinx, the average CS at 10 is 69.0 over all ranked players and 74.3 in Emerald and above. Your opponent improves with you, which keeps the numbers close.

That also means the target you see quoted everywhere, “10 CS per minute in high elo”, is overstated. Master+ Caitlyn players average around 8.5 CS per minute over a whole game.

RoleWhat to measureWhat the data says
ADCCS at 10Jinx averages 69 at 10 minutes over all ranks and 74 in Emerald+. Around 8.5 CS per minute over a game is Master+ level.
Top and midCS at 10, and the difference with your lanerOnly 4 to 9 CS separate all ranked players from Master+ at 10 minutes. The difference with your opponent tells you more than the raw number.
JungleJungle CS at 10 and the difference with the enemy junglerDepends heavily on the champion and on how often you gank; compare with your own games on the same champion.
SupportNot CSSupports earn through their support item and assists. Taking farm from your ADC costs the lane; measure vision instead.

One more caution: patch 26.1 changed the early game. Minions now spawn at 0:30, turret plates stay longer and are worth less after 11 minutes, and lanes got their own role quests. CS benchmarks from before 2026 come from a different game, so don't chase them.

Set your own target

The best benchmark is you. Take your last 20 games in one role, write down CS at 10 for each, and use the middle value (the median) as your starting point. Then aim for five more. Five CS is about one extra wave and a bit, small enough to reach within a couple of weeks and big enough to show in your gold.

If your median is far below the averages above (say, 60 at 10 minutes on an ADC), your farming is the cheapest rank you will ever gain. If you are already at or above them, CS is not your bottleneck; pick a different goal, such as fewer deaths before 14 minutes.

How to improve your CS

1. Make last hits the rule, trades the exception

Most lost CS in real games isn't a missed click; it is a trade taken at the wrong moment, with a minion dying while you were busy. Make it an if-then plan before the game: “If I have to choose between a trade and a last hit, then I take the last hit.” Plans written in this form have a well-documented effect on following through (d = 0.65 across 94 tests).

2. Ten minutes in the practice tool, twice a week

Start a practice game on your main champion with no items or runes beyond the start, and only last-hit for ten minutes. No abilities for the first round, then with the abilities you would use in lane. Count your CS at 10:00. When it beats your ranked median by ten, the gap in ranked is decision-making, not mechanics.

3. Learn the tower pattern for your champion

The hardest CS in lane is under your own tower. Melee and caster minions take different numbers of tower shots before your auto-attack finishes them, and the exact pattern depends on your champion's damage and the patch. Learn it in the practice tool by letting a wave crash into your tower and timing your hits.

4. Recall on a pushed wave

Recalling while a big wave walks into your tower burns those minions. Push first, then back: “If I have the gold for my next item and my wave is pushed, then I recall.” You come back in time for the next wave instead of watching your tower take your gold.

5. Stay alive

Every death costs you the waves that arrive while you are gray, and often the ones that crash into your tower while you walk back. A death that gives your team nothing can easily cost more CS than any last-hitting mistake.

6. Fix the common leaks

Most players lose CS in the same few places. Check your last few games for these:

  • The cannon minion. It is worth more than the others; never let it die to your opponent or the tower without trying.
  • The wave after a kill or a death. Players celebrate or tab out and lose the next wave. Look at the wave first.
  • Walking back from a recall. If you arrive after the wave has crashed into your tower, your recall timing was off, not your last-hitting.
  • Roaming without a pushed wave. A roam that gains nothing costs a whole wave in your lane. Push first, then go.

CS after the laning phase

After the first towers fall, CS per minute is mostly about the side lanes. Waves keep coming in all three lanes whether anyone is there or not. A simple habit: before you join the next group fight, ask whether a side wave is about to crash, and whether you can catch it and still arrive in time. Carries who catch side waves between objectives reach their item spikes earlier, and that is where the difference between 7 and 8 CS per minute over a game comes from.

If you use LazyElo, the overlay shows every player's CS per minute per lane during the game (the same number as TAB, easier to read), and the review after each game compares your CS at 10 with your own average on that champion. Make CS at 10 your goal on the Focus page and it checks every game for you. For a full climbing routine around a goal like this, see how to get better at League.

Questions and answers

Is 7 cs a min good?

Yes, for a laner it is a good number. Master+ Caitlyn players average about 8.5 CS a minute over a whole game, so 7 is not far from high-elo level, and the average ranked Jinx has 69 CS at 10 minutes, just under 7 a minute. Supports aren't meant to farm, and for junglers it depends on the champion.

How much cs at 10 minutes?

Around 70 is a solid target for a laner. Jinx averages 69 CS at 10 minutes over all ranked players and 74 in Emerald and above, and a typical top laner gains only 4 to 9 CS between all ranks and Master+. A better target is your own: the middle value of your last 20 games in that role, plus five.

Sources

  1. Carnegie Mellon statistics capstone (2024), what predicts a win at 10 minutes
  2. League of Graphs, Aatrox top lane stats by rank
  3. League of Graphs, Jinx ADC stats by rank
  4. League of Graphs, Caitlyn ADC stats, Master+
  5. HighGround Gaming, what is a good CS per minute
  6. Riot Games, patch 26.1 notes
  7. Gollwitzer & Sheeran (2006), implementation intentions and goal achievement

Your CS at 10, after every game

LazyElo's review compares your CS at 10 minutes with your own average on that champion, and Focus can make it your goal for the next fifteen games. Both are part of Plus, with 7 days free to try. For Windows, without ads.

  • How to get better at League

    How to get better at League of Legends without just playing more: six habits backed by research and match data.

  • LoL ranks explained

    All 10 League of Legends ranks in order, Iron to Challenger, with divisions and LP.

  • LP gains and MMR

    What MMR is, why no site can check it, why you lose more LP than you gain, and what a dodge costs.

  • KDA explained

    How KDA is calculated in League of Legends, a KDA and kill participation calculator, what a good KDA looks like per role, and why one death fewer beats one kill more..

  • Vision score

    What vision score measures, why the ward difference with the enemy matters more than your own count, what good vision looks like for supports, junglers and laners, and when and where to ward..

  • How to stop tilting

    Tilt is measurable: win rates drop in long sessions.

  • Champion pool

    How many champions to play per role at your rank, the three slots of a good pool, and which champions cover each other's worst matchups on the current patch, from our tier data..

  • Counter picking

    What makes one champion counter another in League of Legends, when a counter pick is worth it, when it isn't, and the hardest matchups of the most played champions on the current patch..

  • Best blind picks

    The safest blind picks for top, jungle, mid, ADC and support on the current patch, from our tier data: high tier and few hard counters.

  • Season dates 2026

    When each 2026 ranked season of League of Legends started and ended, what Riot has said about the end of Season 3, and what resets when a season or a year ends.

  • All guides