⚡The short answer, tier by tier
Game budgets are not mysterious; they are just poorly labelled online. Most "how much does a game cost" articles mix a Flappy Bird clone and a live-service looter into one paragraph. The honest unit of analysis is the tier, because each tier is a different product category with different economics, not just a bigger version of the last one.
The table below is the summary this whole article unpacks. Ranges are what we observe across the market for outsourced or blended-team builds; a fully US-based senior team lands toward the top of each band, a blended North America plus offshore team toward the middle.
Read the "primary cost driver" column carefully. In games, unlike business software, the budget lever is usually content volume — art, levels, animation — not engineering complexity. A buyer who understands that negotiates a completely different, and cheaper, contract.
Our game development services, by tier
| Tier | Typical build cost | Timeline | Team size | Primary cost driver |
|---|---|---|---|---|
| Hyper-casual mobile | $15,000 – $50,000 | 6 – 12 weeks | 3 – 5 | Iteration speed; many prototypes per hit |
| Casual / puzzle | $50,000 – $150,000 | 3 – 6 months | 4 – 8 | Level content volume and polish |
| Mid-core (RPG, strategy, shooter-lite) | $150,000 – $600,000 | 6 – 14 months | 6 – 15 | Systems depth plus art and animation volume |
| AA (premium mobile / PC / console) | $600,000 – $3,000,000+ | 12 – 30 months | 10 – 40 | Art direction, content hours, platform certification |
| Live-ops layer (any live game) | 10 – 25% of build cost per year | Ongoing | 2 – 8 | Content cadence — events, seasons, balance |
🎮What each tier actually buys
Hyper-casual is a portfolio business, not a product business. Nobody budgets $50,000 for one hyper-casual game expecting it to return; the model is to build several simple prototypes, test marketability with cheap ad spend, and kill nine out of ten before soft launch. If a studio quotes you a single hyper-casual title without mentioning the portfolio math, they are selling you a lottery ticket as if it were a product.
Casual and puzzle is the first tier where a single title can be a real business: match-3, merge, word, idle. The build cost is dominated by level content — a puzzle game with 300 levels at launch is a content production pipeline, not just an app — and by the meta-layer (progression, monetization, retention loops) that separates a $150,000 title from a $50,000 one.
Mid-core is where engineering starts to rival art as a cost driver: real systems design, server-side authority for PvP, economy balancing, deeper progression. A mid-core game also crosses the threshold where live-ops stops being optional — players of a $400,000 strategy title expect events and seasons, so the launch budget is not the budget.
AA is a different profession. Twelve to thirty months of production, real art direction, narrative, platform certification for console, and marketing costs that can rival the build. We scope AA honestly: if your budget is under $600,000 and your reference titles shipped with 100-person teams, the answer is to shrink the design, not to find a cheaper quote for the same design.
A note on what sits between tiers, because most real projects land there: a polished casual game with mid-core retention systems, or a mid-core design with hyper-casual art. That is fine — the tiers are planning anchors, not boxes. The way to use them is to identify which tier your content volume and systems depth belong to, price that, and then price the exceptions explicitly. What does not work is averaging two tiers and hoping the number lands in the middle; the expensive half of a design does not get cheaper because the other half is simple.
🎨The art vs engineering split — where the money actually goes
In business software, engineering is 50 to 60 percent of the budget and design is a supporting line. In games, the ratio inverts as you move up the tiers. Art, animation, audio, and level content routinely consume 40 to 55 percent of a mid-core or AA budget, and this surprises buyers coming from app development who expect the programmers to be the expensive part.
The practical consequence: the biggest cost lever in most game projects is content scope, not engineering efficiency. Cutting 100 levels from a puzzle game or two character classes from an RPG saves more than any negotiation on hourly rates. Conversely, "we will figure out the art style as we go" is the single most expensive sentence in game development, because restyling a game mid-production means redoing already-paid art.
The split below is a typical market breakdown by genre — a planning anchor, not a quote. Your split will move with art style: 2D stylized is far cheaper per asset than 3D realistic, and that single decision can halve or double the same design.
| Genre / tier | Art, animation & audio | Engineering | Design, QA & production |
|---|---|---|---|
| Hyper-casual | 25 – 35% | 40 – 50% | 20 – 30% |
| Casual / puzzle | 35 – 45% | 30 – 40% | 20 – 30% |
| Mid-core RPG / strategy | 40 – 50% | 30 – 40% | 15 – 25% |
| Shooter / action (mid-core+) | 45 – 55% | 30 – 40% | 15 – 20% |
| AA narrative / adventure | 45 – 55% | 25 – 35% | 20 – 30% |
The most expensive sentence in game development is "we will figure out the art style as we go." Art is 40 to 55 percent of a mid-core budget, and restyling mid-production means buying it twice. Lock the art direction with a paid vertical slice before committing the full budget — it is the cheapest insurance in the entire project.
⚙️Engine costs: Unity and Unreal licensing, honestly
Engine licensing is a real budget line but a smaller one than the internet argues about. Here is the honest state of the two mainstream choices as of writing — verify at contract time, because both vendors have changed terms in recent years.
Unity: the Personal tier is free for individuals and companies under a revenue and funding threshold (currently $200,000 per year). Above that, Unity Pro is a per-seat subscription, roughly $2,000 per seat per year at list price. For a ten-person team, budget on the order of $15,000 to $25,000 per year in engine seats — meaningful, but single-digit percentages of any mid-core budget. Unity cancelled its proposed per-install runtime fee in 2024 after the industry backlash, so there is currently no per-install charge on new projects built on current terms.
Unreal: free to use for development, with a 5 percent royalty on gross revenue above $1 million per title — and no royalty at all below that line. For a game projected to earn well past $1 million, model the royalty as a real cost of goods; for anything smaller, Unreal is effectively free. Epic has also adjusted royalty terms for Epic Games Store distribution in the past, so distribution channel and engine cost interact.
Godot is genuinely free (MIT licence, no royalties, no seats) and increasingly production-viable for 2D and smaller 3D titles — the trade-off is a smaller hiring pool and fewer middleware options. Custom engines only make sense when the game itself is the technology bet; for anything else, a custom engine adds $200,000 and six months before the game exists.
The engine decision also reprices the team, not just the licence. Unity and Unreal developers are abundant in every market; Unreal C++ specialists bill at the top of the range, and a Godot-heavy stack narrows your future hiring options even though the engine itself is free. When two engines can both ship the game, pick the one with the deeper talent pool in the market your team will actually be hired from — over a multi-year project, staffing flexibility is worth more than any seat price or royalty percentage.
One honest caveat on middleware, because it surprises first-time buyers: engine cost is not tooling cost. Backend services, analytics, crash reporting, multiplayer infrastructure, ad mediation, and build tooling each carry their own subscriptions or usage fees. For a live mid-core title these commonly total $500 to $5,000 per month — a line that belongs in the live-ops budget below, discovered in planning rather than on the first invoice.
| Engine | Cost model | Honest budget impact |
|---|---|---|
| Unity Personal | Free under ~$200K annual revenue/funding | Zero for prototypes and small titles |
| Unity Pro | Per-seat subscription, roughly $2K/seat/yr at list | $15K – $25K/yr for a 10-person team |
| Unreal Engine | 5% royalty on gross above $1M per title | Zero below $1M; model it as COGS above |
| Godot | Free, open source (MIT) | Zero — cost shows up as hiring pool instead |
| Custom engine | N/A | Adds $200K+ and months before gameplay exists |
👥Team composition by genre
Game teams are not generic engineering teams with sprites. Each genre has a composition that experienced studios arrive at because the failure modes of the genre demand it. Use the table below to sanity-check any quote you receive: a mid-core RPG quote with no economy designer or a multiplayer quote with no backend engineer is not cheap, it is incomplete.
A few roles deserve explanation because buyers consistently miss them. The technical artist — the person who sits between art and engineering, owning shaders, pipelines and performance on device — is the highest-leverage hire in 3D production and the one most often cut from budget quotes. The game/economy designer is what separates a mid-core game that monetizes from one that is merely finished. And QA in games is not optional overhead: device fragmentation on Android alone makes a real QA function cheaper than the review-bomb alternative.
Timeline and team size compound: a six-person team for twelve months and a twelve-person team for six months are not the same product. Larger teams buy calendar compression at a real coordination cost — expect 10 to 20 percent overhead per doubling of team size past about eight people.
One staffing decision with outsized cost impact is which roles must be senior. In games, the answer is narrower than buyers assume: the lead engineer, the technical artist, and the game designer carry the project. Mid-level engineers and production artists execute well under real direction. A quote that is senior across every seat is charging you for insurance you do not need; a quote with no senior in those three chairs is not insured at all.
Also budget the producer properly. In software projects buyers sometimes treat project management as overhead to be trimmed; in games the producer is the person who stops content scope from silently doubling between milestones. On a twelve-month mid-core build, a competent producer at 10 to 15 percent of budget routinely saves more than they cost in prevented scope drift alone.
How we staff game teams at Codazz
| Genre / tier | Core team composition | Typical size |
|---|---|---|
| Hyper-casual | 1–2 game engineers, 1 artist, 1 designer, fractional QA | 3 – 5 |
| Casual / puzzle | 2–3 engineers, 2 artists, level designer, QA, fractional PM | 4 – 8 |
| Mid-core RPG / strategy | 3–6 engineers (incl. backend), 3–6 art/animation, economy designer, QA ×2, producer | 8 – 15 |
| Multiplayer shooter | 4–8 engineers (netcode, backend), 4–8 art, technical artist, designer, QA ×2+, producer | 10 – 20 |
| AA narrative / adventure | 6–15 engineers, 8–20 art/cinematics, narrative, audio, QA team, production ×2 | 15 – 40 |
📈Live-ops: the budget that starts after launch
The most common budget failure in games is not the build overrunning — it is the build succeeding and the company discovering there is no money left to operate the game. A live game is a content subscription you owe your players: events, seasons, balance patches, new levels, new characters. Stop paying it and the player counts decay within a quarter or two.
Honest market ranges for ongoing live-ops: a casual title with a monthly event cadence runs $5,000 to $15,000 per month with a small blended team; a mid-core game with seasons and regular content drops runs $15,000 to $50,000 per month; AA live games sit above that. As a planning rule, 10 to 25 percent of the build cost per year, every year the game is live.
Infrastructure belongs in this line too. A multiplayer mid-core title with real concurrency pays $1,000 to $10,000+ per month in backend hosting, scaling with players — the happy problem of success raising your AWS bill should be in the financial model, not a surprise.
The strategic point: decide before launch whether the game is live or finite. A finite premium game — ship it, patch it, move on — is a perfectly respectable model and much cheaper. What does not work is launching a game designed around retention loops and then funding it like a finite one.
| Live-ops line item | Casual title (monthly) | Mid-core title (monthly) |
|---|---|---|
| Content cadence (events, levels, seasons) | $3,000 – $8,000 | $8,000 – $25,000 |
| Balance, patches, community fixes | $1,000 – $4,000 | $4,000 – $12,000 |
| Backend / hosting / tooling | $500 – $3,000 | $1,000 – $10,000+ |
| Analytics, UA support, store management | $500 – $2,000 | $2,000 – $5,000 |
| Total live-ops budget | $5,000 – $15,000 | $15,000 – $50,000+ |
💥What actually blows game budgets
After art restyling — covered above — the killers are consistent across the industry and worth naming before you sign anything.
Multiplayer added late. Real-time networked play is an architectural decision, not a feature. Retrofitting it into a game built as single-player typically costs 50 to 100 percent of the original build. If there is any chance the game needs multiplayer, it must be in the architecture from week one, even if the feature ships later.
Platform certification underestimated. Console certification and storefront compliance are real work packages — weeks of engineering and QA per platform, plus the fee and devkit logistics. Porting "later" to a platform the game was not engineered for is a rewrite in a trench coat.
Soft launch ignored. For free-to-play, the soft launch in a small market is where retention and monetization get measured honestly. Skipping it to save three months means discovering your day-7 retention is broken in front of the whole market instead of one small country. Budget the soft launch, and budget one full iteration cycle after it.
Prototype discipline skipped. For anything mid-core and up, a paid vertical slice — one polished slice of the real game, real art direction, real core loop, four to twelve weeks — is how you de-risk the remaining 90 percent of the budget. It is the game industry equivalent of the discovery phase, and it is the best money in the project.
And the quietest killer: decision latency. Games are thousands of small decisions — feel, pacing, balance, art direction — and every one that waits a week for stakeholder sign-off burns a week of a team that costs five figures a month. A named decision-maker with daily availability is not a management nicety in game production; at mid-core burn rates, slow approvals are a line item larger than the engine licence.
Budget the contingency the same way you would for any software program: 15 to 25 percent held separately, with a stated release mechanism. Games add one twist — hold a portion of that contingency specifically for the post-soft-launch iteration, because the highest-probability use of the money is fixing what real players reveal. A budget that is fully allocated at launch day has no way to respond to the first honest data the game ever produces.
🧭How to buy game development without the horror stories
The procurement shape that works for games: a paid game design document plus vertical slice as phase one, full production as phase two with a go/no-go between them, and a pre-agreed live-ops retainer if the game is a live title. Any studio that wants to jump straight to full production on a mid-core budget is asking you to fund their learning.
Ask for the art pipeline before the engineering plan: how many assets, at what fidelity, produced by whom, and what a restyle would cost. Ask who the technical artist is. Ask what the soft-launch plan is and what metrics gate the global launch. A studio with real shipped titles answers these in specifics; a website with a portfolio of other people’s games answers in adjectives.
We have shipped games across mobile tiers since 2018 as part of 500+ delivered projects, with 200+ engineers and artists between Edmonton and Chandigarh, and the pattern is the same every time: the vertical slice pays for itself, art scope is the budget lever, and live-ops decided before launch is the difference between a game and an expense. If you have a concept and want it scoped into one of these tiers honestly — including being told the concept needs to shrink to fit the budget — that is a conversation we enjoy having.
Scope your game with our teamMobile game development: the full guide