Nobody really buys game art. They buy a game that ships. The art is the part everyone can see, so it gets the credit, but the stuff that decides whether a project lands is almost never the art. It's the pipeline, the communication, the engineering, the discipline around the build. So the question you're actually asking when you go looking for external help isn't "who can make this look good." It's "who can deliver this production."
That framing is usually the miss, so let's get it on the table early: we're a co-development studio. Art is a big part of what we do, but it's one part. We also do animation, VFX, technical art, level design, in-engine cinematics, and real engineering and tools on top of that. We're not selling images. We're selling a team that plugs into yours and delivers work that survives the actual production path.
This is an ad. The honest kind. The advice below is what we'd check if we were hiring us.
What you're actually hiring
What separates a studio from a bunch of artists with a shared folder is what they can actually do. If a partner only does the pretty pictures, the moment you need a shader that holds 30 fps on low-end mobile, a tool to batch out a hundred props in a week, a cinematic that has to play in-engine, or a level pass that hits a performance target, you're back on a call with three more vendors. That's the difference between outsourcing a slice and buying a production partner.
We do all of it. Shipped art, animation and cinematics on big worlds like Diablo 4 and Halo Infinite, VR like Marvel's Deadpool VR, live-service hits like Empires & Puzzles, and tools we built for Riot games. Different budgets, different engines, one bar.
The three dials
Every serious studio is really working with three dials. Quality. Price. Technology.
Quality is the one nobody wants to admit is as much your job as theirs. There's a saying among people who run these teams: you get the quality you enforce, not the quality you describe. A good studio pushes back on the brief. They ask what the frame budget is, whether that asset really needs four LODs, why the prop is in both the environment list and the tech-art list. If they say yes to everything, they're not agreeing with you. They're deferring the argument to production, where it gets expensive.
Price is the warning light. A cheap quote usually has the rework, the delays, and the project management hours baked into it, they just bill those to your calendar. And if a quote lands way below everyone else, ask who's actually doing the work. The A team shows up to the pitch. The B team shows up to production.
Technology is the one that actually contracts time. A studio with real tooling and a real pipeline moves faster, and faster is cheaper without anyone cutting a corner. That's the only kind of cheaper that matters. We build our own tools for the projects we're on. It's the reason a dedicated external team can keep pace with your internal one.
About portfolios
Everyone has a great portfolio. Curation is the whole job of it. So don't ask to see the best work. Ask what shipped and what the studio actually did on it. Good answers are almost annoyingly specific: these characters, that environment, this shader pass. Vague answers mean they weren't the ones doing it.
And at the level you're hiring at, the best work is often invisible, under NDA. We've made promo art for Sony franchises like Spider-Man and The Last of Us, key art for The Elder Scrolls, level design and technical art on Call of Duty. Some of it's public, plenty isn't. That's not a studio hiding something. That's what working with people who don't want it shown looks like.
Dedicated team or per-asset
Per-asset is for closing holes. You need twelve props by March, you order twelve props. But if the goal is shipping a product, a dedicated team wins and it's not close. It works like your own team: your standups, your source control, your quality bar, just without a year of recruiting. That's what retainers are for and what we run. Per-asset keeps you in procurement. Dedicated keeps you in production.
The test task thing
It goes both ways and nobody says so. A paid test says you're serious. A free one says you might be building a spreadsheet of future vendors. The studios you actually want are busy, and they read your test design the same way you read theirs. Everyone's estimating everyone. That's not cynicism, that's just what working with serious people feels like.
Three things we'd walk away from
A studio that agrees with everything. That's the tell, and it covers more ground than any other single warning.
A price far below everyone else.
Loose security. References sent before an NDA, source builds handed around casually, no controlled environment. For the kind of company that hires a studio like ours, security is infrastructure, not paperwork. Ours runs NDA-first by default. How someone treats your confidentiality before you sign is exactly how they'll treat it after.
The short version
Ask what shipped and what they did on it. Send real feedback early and watch what comes back. Prefer a partner who builds their own tools. Buy a dedicated team for production, per-asset only to close gaps. Pay for tests. Walk away from total agreement, magic prices, and sloppy security.