Okay, here is the thing about Valorant in 2026. Everyone is locking in the same five or six agents. Jett, Raze, Omen, Killjoy, Tejo, and whoever Riot dropped most recently. If you queued ranked last month, you saw this lineup countless times. The meta is real. The meta works. But the meta is also leaving a handful of genuinely powerful agents on the bench, and the people picking them up are quietly winning rounds while everyone else fights over the same instalocks.

Why some agents get slept on

The reason agents fall out of favor is usually less about strength and more about how players think about the game. A new agent drops, content creators flood YouTube with broken combo videos, ranked queues fill up with that pick for two weeks, and then the cycle repeats with the next release. Agents that quietly do their job without flashy highlight clips end up dropping in pick rate even when their win rate stays steady.

There is also the rank gap problem. The pro meta and the ranked meta are not the same thing. Plenty of agents look mediocre at the top because coordinated pro teams can punish their weaknesses in high-risk, finely balanced situations. Drop those same agents into Diamond or Plat, and they become genuine carry picks because the average opponent cannot read what they are doing fast enough to react. Looking at agent pick trends from the VCT 2026 scene so far, you see a very different picture than what shows up in your ranked games. The agents pros need are not always the agents you need.

The difference becomes more important than people admit. If you are grinding ranked, using a service like Valorant Boosting to push through a stuck rank, the agents that climb are the ones that punish your specific elo, not the ones winning in Masters Santiago. With that in mind, here are five agents the ladder has decided to ignore. The ladder is wrong.

Deadlock

Sentinel | Released June 2023 | The sentinel that gets stronger every patch

Deadlock has been the easy target of the sentinel role since launch. Killjoy and Cypher dominate pick rates. Vyse showed up in 2024 and immediately ate into whatever ground Deadlock had. By every metric, she is the third or fourth sentinel pick in 2026. 

She is also extremely good, and the gap between her perceived strength and her actual strength is one of the largest in the entire game. Her kit, in practice:

  • Barrier Mesh: A deployable wall that blocks player movement and funnels pushes into pre-aimed kill zones.
  • GravNet: A grenade that roots enemies caught in its area, genuinely difficult to play around without coordinated comms.
  • Sonic Sensor: Concusses pushers automatically when triggered and gives free positional information.
  • Annihilation (ultimate): One of the most disruptive ults in the game when played around defuses or post-plants.

Deadlock has received steady buffs since launch. The agent today is meaningfully stronger than the one that originally launched in 2023. Pro players quietly use her on specific maps where her wall plays a role Killjoy lockdowns or Cypher tripwires cannot fill. In ranked, her ability to shut down a rush single-handedly is wildly underrated, especially in metas where opponents pile into one site looking for quick wins. 

She is also one of the few agents whose newer iterations matter. Recent agent releases like Clove have shifted the sentinel meta in interesting directions, and Deadlock benefits from a meta that punishes coordinated executes more than ever.

Play her if: You want to single-handedly turn off enemy site rushes and never get the credit.

KAY/O

Initiator | Released June 2021 | The most overlooked S-tier pick in the game

KAY/O is genuinely one of the best agents in Valorant. He has been since his introduction. People keep forgetting his prowess, and it is hard to understand why. 

His suppression utility is the only ability in the game that completely disables enemy abilities. In a meta increasingly built around stacking utility (Tejo’s missiles, Omen’s smokes, Killjoy’s lockdown, and Vyse’s traps), having one player who can shut all of that off for ten seconds at a time is absurd. The kit breaks down like this:

  • ZERO/Point: A throwable knife that suppresses every enemy caught in its blast area, disabling abilities entirely.
  • FRAG/ment: One of the most consistent map-control grenades in the game, useful for clearing corners or zoning post-plant.
  • FLASH/drive: Reliable, fast, easy-to-use flashes that work on attack and defense.
  • NULL/cmd (ultimate): Three more knives plus a revive timer that lets teammates bring him back if he goes down.

The reason he gets ignored is that none of this is flashy. KAY/O does not produce highlight clips. His value shows up in round wins, not in 1v5 ace videos. He is the agent that wins the round you should not have won, and nobody notices because the kill cam goes to the duelist who frags out after his suppression has cleared the site.

If you are looking for an initiator who actually carries games at every rank, Riot’s official agent roster is worth a few minutes of reading. The kit is even better than it looks on paper.

Play him if: You want to be the reason your team won, even if nobody on your team realizes it was you.

Yoru

Duelist | Released January 2021 | Reworked September 2021

Yoru has been on lists like these every single year since 2022, and surprisingly, he is still on them. Riot has buffed Yoru repeatedly. They reworked his entire kit. They gave his teleport more range, gave Fakeout the ability to flash on death, and made his ultimate genuinely terrifying. None of it matters. Players still pick Jett, lock Jett, instalock Jett, and then complain about not having a duelist who can actually open sites.

Here is what Yoru actually does in 2026: he is the only duelist in the game whose entire kit is built around deception rather than mobility. Gatecrash is a fake-and-real teleport that lets you set up entries from impossible angles. Fakeout is a clone that flashes when shot, which means anyone trying to clear with information ends up blinded by their own carefulness. Dimensional Drift turns him invisible and untouchable, perfect for late-round re-peeks.

The skill ceiling is brutal. The skill floor is honestly fine. Players just refuse to learn him because Jett is easier. Yoru rewards setup and creativity over raw mechanics, which is exactly why he climbs solo queue better than people give him credit for.

Play him if: You are tired of dying first to the enemy duelist and want to be the one setting traps instead.

Harbor

Controller | Released October 2022 | The smoke agent nobody wants to learn

Harbor is the second-most-controversial controller in Valorant, behind only Astra. Most players take one look at his kit, decide it is too complicated, and lock Omen instead. This is a mistake.

Harbor’s smokes are walls of water. Cascade is a movable wall that breaks line of sight along its entire length. Cove is a dome that protects defuses and plant setups. High Tide is the longest, most flexible smoke wall in the game and can be bent around corners during deployment. His ultimate spawns shockwaves that concuss anyone caught in a wide area. Compared to Omen’s two static smokes, Harbor offers something genuinely different: dynamic, shapeable, time-bought map control that punishes any team trying to push through unprepared.

The reason he gets ignored is the learning curve. Walls are not point-and-click smokes. You need to know map angles and timing in a way Omen does not require. But once you spend a few hours in custom games figuring out lineups, Harbor becomes one of the most flexible controllers in the game, especially on attacker-sided maps where wall smokes shape entire executes.

Looking at competitive viewership data, Esports Charts tracks Valorant pick trends across regions, and Harbor consistently appears in execute-heavy comps from Pacific teams that the rest of the world is still figuring out.

Play him if: You like controllers but want more options than two static smokes for an entire round.

Iso

Duelist | Released October 2023 | The 1v1 specialist nobody learns

Iso came out in late 2023, had a brief moment of being the best duelist in the game, got nerfed, and then completely dropped off the map. The post-nerf version of Iso is still very strong. Players just gave up on him.

His kit is built around dueling. Double Tap creates a shield bubble that absorbs one bullet, which sounds underwhelming until you realize how often duels in Valorant come down to a single shot. Contingency creates a wall that blocks bullets but lets him push through. Undercut is a fast vulnerability tag. Kill Contract, his ultimate, drops him into an arena with one specific enemy in a forced 1v1, with no abilities and no help. Win it, and the round dynamics change instantly.

The thing about Iso is that his entire identity is the duel. He is not a fast-entry duelist like Jett or Raze. He is not a flashy aim-check duelist like Phoenix. He is the duelist who walks into a 50/50 and turns it into a 70/30 because of one absorbing bullet shield. That sounds boring on paper. In practice, it is the most consistently round-winning duelist kit in the game outside of the top instalocks.

According to Liquipedia’s Valorant agent and tournament data, Iso has hovered near the bottom of duelist pick rates for over a year despite stable performance numbers. This is the gap. This area is where free elo lives.

Play him if: You like dueling and you want a kit that makes every duel slightly less of a coinflip.

The bigger picture: why the meta keeps missing

Step back, and a pattern emerges. The agents on this list share one common trait: they all reward thinking over reflexes. The archetypes break down like this:

  • Yoru: deception specialist who turns information against the enemy.
  • KAY/O: a utility shutdown initiator who breaks the abilities meta itself.
  • Harbor: dynamic map control built around walls instead of smokes.
  • Deadlock: anti-rush sentinel whose value scales with how aggressively opponents play.
  • Iso: duel engineer who tilts every coinflip in his favor.

The Valorant ranked meta has steadily drifted toward agents that produce instantly measurable impact: rapid entries, immediate kills, and highlight clips. Agents that produce slower, harder-to-measure value end up underrated even when their win rates suggest otherwise. The pro scene is starting to figure this aspect out. Pacific teams have used Harbor and Deadlock in compositions that the rest of the world still considers off-meta. Players who learned Yoru and KAY/O years ago have built individual climb stories on agents most people refused to consider.

The lesson is not that the meta is wrong. The meta is right for what it measures. The lesson is that the meta might be measuring the wrong thing if your goal is to actually win more games at your specific rank. The agents that climb are the ones that punish the way average players make mistakes, not the ones that win on a perfectly executed pro stage.

Five agents. Five chances to break the losing streak you keep queuing into. Pick one. Spend a week on them. Stop defaulting to the same instalock that hasn’t won you a game since the last act. Your rank will reflect it eventually.