In ranked Valorant, your agent pick is your opening statement to the lobby. It tells your team what you’re building, which angles you’ll cover, and where you expect to create value. Pick the wrong agent for the map or the composition, and you start the game asking teammates to compensate before a single bullet is fired.
This tier list covers Act 2 of Season 2026, based on win rate and pick rate data updated April 28, 2026. The current competitive pool runs seven maps — Abyss, Bind, Breeze, Corrode, Haven, Pearl, and Split — established with Patch 12.00 in January. Two new agents joined this act: Miks (Controller) and Waylay (Duelist). All rankings reflect ranked matchmaking performance, not pro play.
If you’re building foundational knowledge before diving into tier placements, our Valorant Agent Cheat Sheet covers all 29 agents with role descriptions and ability summaries. The Valorant Beginner’s Guide 2026 walks through economy rules and the fastest path to reaching ranked.
Full Agent Tier Table — Act 2, Season 2026
Data source: MetaBot, updated April 28, 2026. Win rates reflect ranked matches across the current 7-map pool. Verified on Patch 12.05.
| Tier | Agent | Role | Win Rate | Pick Rate |
|---|---|---|---|---|
| A | Neon | Duelist | 53.7% | 5.5% |
| A | Clove | Controller | 53.6% | 14.6% |
| A | Vyse | Sentinel | 53.1% | 1.1% |
| A | Skye | Initiator | 52.9% | 2.8% |
| A | Tejo | Initiator | 52.9% | 0.7% |
| A | Fade | Initiator | 52.5% | 5.8% |
| A | Sova | Initiator | 52.1% | 7.8% |
| A | Jett | Duelist | 52.0% | 11.6% |
| B | Cypher, Phoenix, Viper, Raze, Chamber, Killjoy, Sage, Reyna, Brimstone, Deadlock, Yoru, Waylay | Mixed | ~51.2% avg | Varies |
| C | Iso, Veto, Miks | Mixed | ~49.3% avg | Varies |
| D | Omen, Harbor, KAY/O, Astra, Breach, Gekko | Mixed | ~46.5% avg | Varies |
No agents reach S-tier by MetaBot’s methodology this act — the A-tier eight form the true ranked ceiling. Worth noting upfront: different sources split on Clove and Sage. PCGamesN lists both as S-tier. Win-rate data places them in the top half of A-tier. The practical difference is whether you call 53.6% a floor or a ceiling — either way, both are correct ranked picks.
Controllers: Queue Clove. Don’t Touch Omen.
Clove’s 53.6% win rate is the second highest across all roles, and her 14.6% pick rate is the highest in the entire A-tier. That combination — high win rate and high pick rate — is the clearest signal in the data. The Valorant player base has voted with their queue, and Clove is winning those games above average across the board.
The mechanism that separates Clove from every other controller isn’t the smokes — it’s that she stays relevant after dying. Traditional controllers go down and leave their team with no coverage. Clove goes down and can still swing a round if a teammate creates a pick opportunity, because her passive lets her reactivate smokes from death and her ultimate (Not Dead Yet) triggers a revive on kill. PCGamesN calls her “a Controller in title only” — the dueling upside and post-death utility sit genuinely outside what the role has historically provided.
The counterintuitive pick to address: Omen sits in D-tier at roughly 46.5% win rate average. His map-specific data makes this concrete. Across the current 7-map pool, Omen exceeds 50% win rate on exactly one map — Corrode at 51.6%. On the six remaining maps in ranked rotation, his win rate ranges from 44.1% on Bind to 49.1% on Haven. If your ranked habits include Omen because teleport tricks feel oppressive, the data says those games are costing you more wins than they’re generating.
When NOT to pick Clove: If your lobby already has two duelists taking uncoordinated duels, adding smokes doesn’t fix the fundamental problem — no one is trading efficiently. In that team state, a second sentinel and methodical half-round play wins more rounds than creative smoke usage ever will.
Initiators: Four A-Tier Agents. Run Two.
The initiator role has the most A-tier agents this act — Skye, Tejo, Fade, and Sova all sit above 52% win rate. It also has the widest performance gap between top picks and the next tier down. This isn’t random. The current 7-map pool rewards information-gathering before committing to a site. Teams running two initiators gain twice the pre-push intel, which translates directly to winning more gunfights in positions the enemy hasn’t pre-aimed.
Sova’s map-specific data is the clearest individual case study available this act. On 5 of the 7 maps in the current ranked pool, his win rate exceeds 50%:
| Map | Sova Win Rate | Verdict |
|---|---|---|
| Split | 62.7% | Best map — bolt angles clear 8+ positions simultaneously |
| Corrode | 53.9% | Strong — open corridors reward drone coverage |
| Breeze | 53.0% | Long sightlines extend bolt utility |
| Abyss | 52.2% | Consistent across both sites |
| Haven | 51.3% | Three-site layout creates drone value on C-Long |
| Pearl | 49.7% | Below 50% — corner geometry limits bolt angles |
| Bind | 44.9% | Avoid — queue Fade here instead |
Split’s 62.7% is one of the highest single-map win rates for any agent in the current pool. The mechanism: Split’s sightlines and vertical geometry create exact bolt angles where Sova’s recon bolt reveals multiple enemy positions simultaneously before committing entry. That’s not a general map-fit statement — it’s a specific mechanic that rewards practiced bolt lineups with a 62.7% win rate payoff. If you’re grinding Split, Sova is the initiator.
On Bind, TP-heavy rotation speed and corner-based geometry break Sova’s bolt angles — the TP forces fast rotations that make drone intel stale before the team can act on it. Fade is the direct swap. Her Haunt ability reveals enemies through terrain without a lineup requirement, and her Seize debuff scales with game sense rather than memorized positions. Fade’s 52.5% win rate holds across map types because of that adaptability.
Skye’s 52.9% win rate with a 2.8% pick rate makes her the most underplayed A-tier agent this act. Her Guiding Light flash combined with Trailblazer generates more information per ability use than any other initiator when a teammate is prepared to follow. The low pick rate reflects a solo-queue coordination problem — she needs a teammate to commit to the flash window — not a kit problem. If you have a regular partner or play in a trio, Skye is worth the investment.
Tejo (52.9%, 0.7% pick rate) is statistically tied with Skye but running on a very small sample. A 0.7% pick rate means the win rate reflects mostly experienced or experimental players. Hold off on prioritizing him for ranked until more data accumulates across a broader player pool.
When NOT to pick Sova: Bind and Pearl. On Pearl, close-corner fighting neutralizes bolt angles before the team can act on the reveal. On Bind, the pick is Fade. On every other map in the current pool, Sova is the default correct initiator selection.
Sentinels: Vyse Leads, Cypher for Bind and Haven
Vyse tops the sentinel role with a 53.1% win rate — third highest across all 29 agents this act. Her sample size is smaller than Clove or Sova (1.1% pick rate), but the gap from B-tier sentinels is structurally consistent. After the Steel Garden area restoration in Patch 12.00, her site-approach denial forces attacking teams to burn utility just to create a push window — which is exactly what a sentinel is supposed to do.
Cypher sits in B-tier, but his value is map-specific. On Bind’s hookah corner and Haven’s B-short, Cypher’s camera network generates information density that Vyse’s lockdown kit doesn’t match. If you know these maps well and your information-gathering instinct is developed, Cypher is the correct pick on those two maps specifically. Outside Bind and Haven, Vyse is the better default sentinel.
Killjoy (B-tier) is the correct starting point if you’re learning the sentinel role below Gold. Her turret and Nanoswarm setup is predictable enough that experienced players clear it — but at lower rank brackets, clearing turret + grenades requires utility spend that still benefits your economy. The lower skill floor means you execute the sentinel role correctly from game one, without needing deep setup knowledge.
When NOT to pick a second sentinel: If your composition has no duelist, two sentinels creates a team that holds position but can’t take ground. Sentinels close sites; duelists open them. Without an opener, you will consistently lose pistol rounds and anti-eco rounds where defensive holds don’t compensate for lost space.
Duelists: Neon Wins the Data. Jett Wins on Mechanics.
Neon’s 53.7% win rate is the highest of any single agent in A-tier this act. The reason is mechanical rather than situational: Sprint and Slide don’t just generate speed — they reset angle exposure. When Neon slides behind cover during a fight, she’s repositioning to take the same angle with first-shot advantage on the next peek. Players who build this movement habit into their game win more 50/50 duels across a full session than any other duelist kit provides.
Jett’s 52.0% win rate with 11.6% pick rate makes her the second most-picked agent in A-tier, behind only Clove. The gap between Jett’s ceiling and her floor is wider than any other duelist. Practiced updraft-dash mechanics create pressures that no other agent can replicate: repositioning mid-duel, holding aggressive Operator angles, and disengaging from unfavorable fights before they resolve. At lower ranks, those mechanics are underdeveloped — which is why her win rate sits 1.7 points below Neon’s despite the higher pick rate.
The practical pick guide: if your strength is consistent aim and your dueling is proactive (you create fights), Neon. If your strength is holding creative angles and playing for picks (you wait for fights to come to you), Jett. Both are correct depending on your expression of the duelist role — don’t switch because a tier list percentage difference looks meaningful in isolation.
Raze (B-tier) still clears corners with satchel-launched grenades, but post-nerf she no longer functions as an explosive team opener. She’s viable if your playstyle centers on the grenade corner-clear mechanic specifically. Otherwise, Neon or Jett are the correct A-tier picks.
When NOT to pick Neon: If your aim consistency doesn’t let you exploit the exposure windows her speed creates, movement creates vulnerability without payoff. Jett’s dash is reactive — you deploy it responding to a developing fight, not anticipating one. That timing is more forgiving while you build out positional instincts.
New Agents This Act: Miks and Waylay
Both agents joined this act and are performing below the A-tier threshold after a full act of ranked data. This is useful context before investing time in either for competitive climbing.
Waylay (B-tier, 50.2% win rate, 5.8% pick rate) is a Duelist built around her Refract ability — a safety beacon that returns her to a prior position while invulnerable. The problem is a structural one: her defense win rate is 51.4% while her attack win rate is 48.7%, a 2.7-point defensive advantage. Duelists are supposed to open sites on attack; Waylay’s kit functions better holding them. In a coordinated team where an initiator is consistently generating pre-push information, Waylay can capitalize effectively on the right angles. In solo queue, that information consistency is too low to unlock her kit reliably. Neon or Jett remain the safer investment.
Miks (C-tier, 49.2% win rate, 0.3% pick rate) is the first controller with a healing ability, but the 1.90 KDA above her role average (1.59) suggests Miks players are fighting more than supporting — which may reflect a learning curve mismatch with a kit designed for team-first play. Her M-pulse toggle (switch between concussion and healing output) demands a mid-fight decision at a moment when most players’ attention is on crosshair placement. A sub-50% win rate after a full act means the kit hasn’t found its reliable ranked application yet. Clove remains the clear controller pick. Check back when Riot adjusts her numbers or the player base develops consistent lineups around her.
The Ranked Composition Formula
Standard competitive composition: 1 Controller + 1 Sentinel + 2 Initiators + 1 Duelist. The two-initiator setup is the specific adjustment the current 7-map pool rewards. Every map in the current rotation has at least one site approach that requires corner-clearing or information before committing — teams with two initiators generate that information twice as consistently before each round decides.
Decision tree for filling your role:
- Map is Split, Corrode, Breeze, Abyss, or Haven: Sova is the initiator pick.
- Map is Bind or Pearl: Fade is the initiator pick.
- No controller in lobby yet: You play Clove, regardless of your comfort pick.
- Team lacks a second info agent: Skye as second initiator beats adding a second duelist.
- Below Gold, learning sentinel role: Killjoy over Cypher or Vyse until setup instinct is developed.
For how agent picks interact with round economy decisions — when to full-buy versus save and how role assignments affect eco rounds — see our Valorant Economy Cheat Sheet.
By Player Type: Same Role, Different Agent
| Player Type | Controller | Initiator | Sentinel | Duelist |
|---|---|---|---|---|
| Learning ranked | Clove | Fade | Killjoy | Jett |
| Ranked climber (Gold–Diamond) | Clove | Sova / Fade (by map) | Vyse | Neon |
| Competitive / Immortal+ | Clove | Skye + Sova | Cypher (map-dependent) | Jett |
Clove is the correct controller at every rank. The differentiator between player types is whether your team can coordinate around her revive-aggressive plays — that coordination develops with experience, not agent switching.
FAQ
Is Omen playable in ranked 2026?
On Corrode specifically — yes. His 51.6% win rate there makes him a neutral-to-positive pick on that map. On every other current-pool map, his win rate falls between 44.1% and 49.1%, which is a structural below-average performance, not a skill-expression problem. Below Platinum, comfort and mechanical familiarity can offset tier disadvantage enough to not matter. Above Diamond, the cumulative gap between Clove and Omen across a session becomes large enough to show in your rank progression. The switch is worth making.
Which initiator should a beginner learn first?
Fade. Her Haunt ability places a visible orb that reveals enemies on detection — no lineup knowledge required, and the reveal is clear enough to act on immediately after landing. Sova’s recon bolt generates more information at high skill levels but requires map-specific lineups that take hours to build. Skye requires coordinated teammate follow-through that solo queue rarely provides consistently. Fade gives useful information from the first game played, on any map in the pool.
Should I main Waylay instead of Neon or Jett?
Not for solo queue ranked climbing. Waylay’s kit needs pre-push information from teammates to create value — her Refract repositioning is most effective when you already know which angle the enemy is holding, which usually comes from an initiator’s prior reveal. In solo queue, that information consistency is too low. Neon creates her own advantages through movement; her value doesn’t depend on teammate inputs. Waylay has a place in premade compositions with a dedicated info-running initiator. Until you have that setup reliably, Neon or Jett climb faster.
Sources
- MetaBot — Valorant Agent Tier List (updated April 28, 2026)
- MetaBot — Sova Best Maps in Valorant 2026
- MetaBot — Omen Best Maps in Valorant 2026
- PCGamesN — Valorant Tier List Season 2026 Act 2 (April 20, 2026)
- MetaBot — Waylay Guide 2026
- MetaBot — Miks Guide 2026
- Turbosmurfs — Valorant Competitive Map Rotation 2026
I've been playing video games for over 20 years, spanning everything from early PC titles to modern open-world games. I started Switchblade Gaming to publish the kind of accurate, well-researched guides I always wanted to find — built on primary sources, tested in-game, and kept up to date after patches. I currently focus on Minecraft and Pokémon GO.