Genshin Impact Character Roster: Elements, Regions and Building Your First Teams

The Roster at a Glance: About 120 Units and Counting

When Genshin Impact launched in September 2020 it shipped with 23 characters. By the time Version 7.0, titled Everwinter Without Mercy, arrived on August 12, 2026, the playable cast had ballooned to roughly 120 units. That number climbs a little every six weeks, so treat it as a living total rather than a fixed one.

What makes the roster interesting is not just the headcount but the variety of roles. Each character is tied to one of seven elements and wields a specific weapon class, and most are designed around a particular combat identity. Some are straightforward on-field damage dealers. Others exist purely to enable a teammate's damage through buffs, healing, shielding, or off-field application of an element. A handful, such as the Traveler, even shift element entirely as you progress the story, effectively functioning as several characters in one.

For a new or returning player, the practical takeaway is simple: you do not need most of these characters to enjoy the game.

Seven Elements Define Every Kit

Every character resonates with exactly one of seven elements. Understanding them is the foundation of all team building:

• Pyro (fire)
• Hydro (water)
• Electro (lightning)
• Anemo (wind)
• Cryo (ice)
• Dendro (nature)
• Geo (earth)

Element does two things. First, it decides which elemental reactions a character can trigger when their element meets another on an enemy. Second, it strongly hints at how the character wants to be played. A Pyro applicator pairs naturally with Hydro for Vaporize or with Cryo for Melt, while a Dendro unit wants to meet Hydro, Electro, or Pyro to set off Bloom, Aggravate, or Burning.

Version 7.0 added a fresh layer to this system in Snezhnaya with the Stellar Glimmer reaction family. The newest member, Stellar Swirl, triggers when Anemo meets Cryo under the right conditions and spawns a Stellar Vortex that detonates for combined wind and ice damage. Even with these additions, the core truth holds: reactions are the engine of the game's damage, not the raw stats of any single unit.

Eight Nations, One Connected Map

Teyvat is built from eight major regions, each tied to one of the seven elements and its own cultural flavor:

Snezhnaya debuted in Version 7.0 as the long-teased seventh nation, a snowbound land of train-connected settlements and the capital city Snezhnograd. New players should not feel pressure to own characters from every region. Any region's units can clear all current content, because strength comes from synergy and artifacts, not geography.

How Characters Enter Your Account

Almost every character is acquired through the Wish system using Primogems, the game's premium currency, which are converted into Fates at a rate of 160 Primogems per pull. There are several banner types:

• Character Event Wish: the limited banner featuring one rate-up 5-star plus three 4-star units, running in two phases per version.
• Weapon Event Wish: limited signature weapons.
• Chronicled Wish: a curated pool of older regional characters and weapons.
• Standard Wish: a permanent pool with a fixed set of 5-star units.
• Beginner's Wish: a one-time discounted new-player banner.

The Traveler is the one major exception: you unlock their elements for free by resonating at each nation's Statue of the Seven. In Snezhnaya, touching the local statues grants Cryo, and Alyosha, a 4-star Electro polearm hunter, is obtainable at no cost by advancing the new Archon Quest.

Regional Banners Run on a Rough Six-Week Clock

A full version lasts about six weeks and is split into two phases of roughly three weeks each. Each phase typically debuts or reruns one or two 5-star characters.

This cadence is the backbone of pull planning. Because a new limited character appears so frequently, you must decide deliberately which few units deserve your Primogems. Chasing every banner is the fastest way to end a patch with nothing built. The gacha math that governs these banners is worth knowing precisely:

• Hard pity on character banners is 90 pulls, meaning a 5-star is guaranteed at or before pull 90.
• The 50/50 gives you a 50 percent chance the 5-star is the featured unit; lose it and the next 5-star is guaranteed to be featured.
• Starting with Version 5.0, Capturing Radiance quietly lifts the effective win rate to about 55 percent over time.
• The worst case to lock a specific character is 180 pulls, equal to 28,800 Primogems at 160 each.

Reactions Matter More Than Single-Unit Power

The single most common beginner mistake is hoarding Primogems for the flashiest individual character and then throwing them into a team with no elemental cohesion. Genshin does not reward that. A modest team whose elements chain into strong reactions will outperform a ragtag pile of premium units that never trigger anything.

The point is not to memorize every multiplier, but to ensure your four slots actively talk to one another.

Assembling Your First Two Teams

Endgame content such as the Spiral Abyss expects two full teams of four, so the goal is to build two functioning squads rather than one perfect one. A reliable skeleton for each team looks like this:

1. Main DPS: the character who stays on field dealing most of the damage.
2. Sub-DPS: a unit who deals meaningful damage off-field or during quick swaps.
3. Support: healing, shielding, or a damage buff that enables the carry.
4. Flex: a second support, a second sub-DPS, or an elemental applicator to complete the reaction chain.

Match elements on purpose. If your main DPS is Pyro, surround them with Hydro and Anemo to enable Vaporize and Swirl. If they are Cryo, pair them with Pyro for Melt. You do not need five-star units in every slot; many 4-star characters are cornerstones of the strongest teams in the game.

Artifacts Beat Levels

A tempting trap for new players is pushing character levels as high as possible while ignoring artifacts. In practice, a character at a reasonable level with a properly built artifact set will deal far more damage than a max-level character wearing random gear. Artifacts are the gear system: five slots per character, each rolling main stats and substats, and full sets granting powerful two-piece and four-piece bonuses.

Prioritize getting a character's artifacts to a sensible level (roughly plus 16 to plus 20 on the key pieces) and matching the set bonuses to their role before chasing tiny level upgrades. Elemental mastery, crit rate, crit damage, and the relevant attack or elemental damage stats are usually what separate a usable unit from a great one.

A Starter Roadmap

If you are beginning today, do not stress about the 120-unit count. Follow this order: finish the early Archon Quests to unlock regions and the free Traveler elements; build one reaction-based team around whatever strong 4-star units you pull; start a second team once the first is comfortable; and only then spend Primogems on the limited characters that fill the holes your roster has. The roster's size is a feature, not a requirement. Play the elements, not the numbers.

Energy Recharge Is a Rotation Problem, Not a Stat Problem

New players treat Energy Recharge as a stat they grudgingly stack on supports. Experienced players treat it as a scheduling constraint: a burst that is not available when the rotation reaches that character is simply not part of the rotation at all, regardless of how much damage it could theoretically deal.

The practical method is to build backward from the rotation. Write out the sequence you intend to run — who casts, in what order — and note every burst you expect to use each cycle. Then ask, for each character, whether their burst is available every single time that slot comes around. Any character whose answer is "sometimes" needs more Energy Recharge before they need anything else, even if their raw damage is high. A marginally lower-damage burst fired every rotation beats a bigger one fired every other rotation.

Off-field characters suffer most from this because they are often not on the field long enough to catch their own energy from attacks, making them dependent on particles generated by teammates. Funneling is the fix: use the energy-generating skill of one character and immediately swap to the character who needs that energy, since particles travel to whoever is currently active. Elements that generate particles generously from their skills, along with characters who group enemies tightly, are the usual enablers here.

The test is straightforward and takes one minute: run your rotation against a boss three times in a row without pausing. If any burst fails to come off cooldown during even one cycle, the build is not finished — regardless of what its crit ratio looks like.

Horizontal Versus Vertical Investment: The Decision Rule

Every player eventually hits the same choice: pull another new character, or invest in constellations and signature weapons for one you already own? There is a clear structural answer, and it has little to do with which characters you personally like.

New characters add capability. Constellations add efficiency. When your account lacks a specific function, no amount of vertical investment will substitute for it, because efficiency multipliers only apply to things your account can already do.

The common failure is doing both halfway: pulling for one copy of everything and then spending whatever is left on a constellation or two. That path produces a wide but shallow roster where nothing is fully built. Choose one axis per planning cycle. If you are going wide, genuinely commit to building each new unit before pulling the next. If you are going deep, commit to finishing the constellation line rather than stopping at an arbitrary midpoint where the upgrade is incomplete.

Signature weapons deserve the same scrutiny. A weapon attached to a character you field constantly returns value every day; a weapon chasing a marginal upgrade while your artefact sets are unfinished is usually the worse purchase, because artefacts are the larger and cheaper multiplier at that stage.

Three Artefact Mistakes and What Replaces Them

Artefact optimization is where accounts quietly waste months. The three highest-cost errors are procedural, not statistical.

The companion habit is faster than any of them: pick one character at a time as your project. Finish their set to a point where you would not be embarrassed using them in difficult content, and only then start the next project. Sequential finishing beats parallel dabbling, and it makes each incremental improvement visible instead of diluted across five half-built characters.

A useful self-check when judging whether a piece is worth keeping: count how many of its substat rolls landed on stats that character actually uses. A piece where two or fewer rolls are useful is a placeholder, not an investment, even if it started with a promising set of lines.

Reading Endgame Chambers as Two Separate Problems

Content requiring two teams is rarely two equally weighted halves. Almost every such floor has one side that asks for damage against many enemies at once and one side that asks for damage against a single durable target, or one side that punishes a specific element and one side that rewards a different one. Treating them as a single "my teams are too weak" problem is why players keep rebuilding teams that were never the issue.

Decompose before you build. Read the enemy list for each half, then classify it: clustered and numerous, or few and durable? Does any enemy apply an element to itself, effectively giving you a free reaction partner? Does anything have a shield that must be broken with a specific element before it can be hurt efficiently? Answering those three questions usually determines the two teams before you consider a single character name.

The second-order consideration is rotation length. A team whose full cycle takes a long time may be perfect against a persistent single target and terrible against a group that must be cleared within a tight window. If your clearing fails on time rather than on deaths, the problem is almost always AoE coverage or cycle time, not raw damage output.

Run this diagnostic whenever you stall: record which half you fail, note whether you run out of time or lose characters, and check which of those mismatches applies. Timed-out failures point to coverage; death failures point to sustain or mitigation. Guessing at both at once is how players spend months rebuilding teams that were solving the wrong problem.