New AI Models, Streamlined Animation, and Your New Home Base
New AI Models, Streamlined Animation, and Your New Home Base
TL;DR:
- A refreshed AI lineup: MiniMax M3, Kimi K2.7, and GLM-5.2, with our two newest models now free for everyone
- A sharper image model: GPT Image 2 is the new default for art generation, and a major step up in image quality
- See what the AI is doing: a live status bar shows the agent's current action, elapsed time, and token count while it works
- Animation, no characters required: go straight from concept art to animation, with reference sheets as an optional step
- More control over your animations: choose the video model, resolution, and clip length, and combine multiple reference sheets in one animation
- One streamlined view: generate, edit, and bake a sprite sheet without re-extracting frames or losing your edits
- A new home base: land on your most recent work and jump straight back in
- Plus a round of fixes and quality-of-life improvements
Smarter Models in Code Studio
The AI behind Code Studio chat reads your project, writes code, fixes bugs, and wires up game logic. Not every task needs the same horsepower, though. A quick tweak shouldn't cost as much as a tricky refactor, and a hard problem deserves the strongest reasoning you can get.
So we've refreshed the entire lineup, and you pick the model that fits the job right from the chat input.
The new lineup
| Model | Cost | Available from |
|---|---|---|
| MiniMax M3 | 1 credit | Free |
| Kimi K2.7 | 2 credits | Free |
| GLM-5.2 | 5 credits | Hobbyist |
The headline: our two newest models, MiniMax M3 and Kimi K2.7, are now available on every plan, including Free.
- MiniMax M3 is the new default. It's fast, balanced, and handles the vast majority of everyday building at just 1 credit per message.
- Kimi K2.7 steps up the reasoning for trickier, multi-step work like planning a feature, untangling a bug, or thinking through architecture, at 2 credits.
- GLM-5.2 brings premium intelligence for your hardest problems.
The model selector shows every option with its quality and per-message credit cost. Models above your tier still appear, locked, with an upgrade prompt, so you always know what the next plan unlocks.
See exactly what the AI is doing
Our coding agent works quietly and just gets things done. But "quiet" used to mean you couldn't tell what was happening. Now there's a live status bar right above the chat input that keeps you in the loop the whole time it's working:
- Its current action, like "Editing GameScene.ts," "Running tests," or "Installing dependencies," pulled from what the AI is actually doing rather than guessed
- An elapsed timer showing how long the turn has been running
- A running token count of the AI's response so you can see it building in real time
The conversation itself is calmer, too. Tool calls collapse into tidy summary rows, the AI's reasoning is grouped into readable steps, and its reply streams in as it's written. And if a turn ever stalls on a dropped connection, a Retry button reconnects right where it left off, with no lost context.
Sharper Art with GPT Image 2
Great animation starts with great art, so we upgraded the model that makes it. GPT Image 2 is our newest image model and a major step up in image quality, and it's now the default behind everything you generate, from concept art and game assets to reference sheets.
There's nothing to switch on. New art uses GPT Image 2 automatically, with quality levels you can choose to balance detail, speed, and cost.
Animation, All in One Place
Animating something used to mean following a fixed path. You started with concept art, turned it into a character, generated a reference sheet, and only then could you animate. Every step was required, even when you didn't need it.
That's gone. There are no characters standing in the way anymore. Your concept art, reference images, and animations now live together in one place, and you take the shortest path that fits:
- Concept art straight to animation
- Or concept art to a reference sheet to animation, when you want multiple angles guiding the motion
If a quick animation is all you're after, you're one step away from it.
Reference sheets, right where you need them
Reference sheets aren't new, but they're no longer locked inside the old character flow. They live in your collection now, next to your concept art and animations. Generate one from any image to get a clean front, side, and back view of the same subject, then reuse it across as many animations as you like. Pin one to an animation and the AI sees the subject from every angle, so the motion stays true to your design.
You're not limited to a single sheet, either. Generate several reference sheets from the same piece of concept art when you want different poses, outfits, or variations to pull from. And you can bring more than one into a single animation. Combine multiple sheets, or mix concept art with a reference sheet, and the AI draws on all of them at once.
Choose your video model
Every animation now gives you control over how it renders. Pick the video model quality that fits the moment: a faster, lighter pass when you're exploring, or the full-quality model when you're locking something in. Set the output resolution to 1080p or 720p, choose your clip length, and decide how many variations to generate at once. What's available depends on your plan.
One view to edit and bake
The animation editor worked before, but it was spread across screens and slow to fiddle with. We pulled it into a single view and focused on making it quick to use:
- Reorder, hide, retime, flip, or duplicate frames in one place, and mirror a sequence into a seamless loop
- Resize and resample without re-extracting your frames from scratch
- Reset your edits fast when you want to start over
- Bake a game-ready sprite sheet, WebP or PNG, with the frame metadata your engine needs
Free accounts get a Lite preview at 720p, 4 seconds, and one video, which is great for trying it out. Paid plans unlock full sprite animation up to 1080p, 8 seconds, and four variations per generation, plus off-platform export of everything you make.
Your New Home Base
Logging in used to drop you on a generic screen. Now you land on a home base built around what you're actually doing.
- New here? You'll get a gentle starting point: spin up your first game, start your first art collection, or take a quick tour of a studio.
- Coming back? You'll see Jump back in, a single grid of your most recent work across both studios, Code Studio projects and Art Studio collections together, sorted by what you opened last, each with a thumbnail and a quiet "2m ago"-style timestamp.
From there it's one click back into anything. Quick links jump straight to all your projects or all your collections, and a New game / New collection tile is always right there when inspiration strikes. No hunting, no guessing where you left off.
Fixes & Improvements
Fixes:
- Your credit balance now updates the instant a generation finishes, so there's no refreshing to see where you stand
- More reliable background removal for props and assets
- Redesigned Studio header with clearer, two-line navigation between tools
Quality of life:
- Animation prompts now accept up to 2,000 characters, so you can give the AI much more detailed direction
- Collections load faster now that we serve a single baked preview image per group instead of many full-size ones
- New members get a starting credit bonus to explore right away
- The in-browser editor now type-checks your code continuously as you build, so you catch TypeScript errors the moment they appear
What's Next
We're going to keep sharpening the AI, the animation tools, and the moment-to-moment feel of building a game. Your feedback drives all of it, so keep it coming.
- The Makko Team
š§ Questions? Reach us at support@makko.ai š¬ Join the conversation on Discord