正在加载视频...

视频加载失败

Introducing MIRA. A playable, multiplayer world model. A dream of Rocket League. Trained on 10k hours of data collected with publicly available bots, MIRA learns the dynamics of a four-player game. The model runs in real time at 20 fps, based on the keys you and the other players...

507,564 次观看 • 3 个月前 •via X (Twitter)

35 条评论

kache 的头像
kache3 个月前

@ricklamers You probably don't want to predict the entire world

General Intuition 的头像
General Intuition3 个月前

Play the four-player demo now.

Palmer 的头像
Palmer3 个月前

10k hours sounds like a lot then I remember some people got like 5x that by themselves lmao

Vincent Micheli 的头像
Vincent Micheli3 个月前

Come hang out with us at our ICML booth to play some matches and learn more about MIRA!

0xobsessed 的头像
0xobsessed3 个月前

i don’t understand what you’re building

paula wehmeyer 的头像
paula wehmeyer3 个月前

Play MIRA at ICML in Seoul with us. Meet the team at booth 111.

Aakash Reddy 的头像
Aakash Reddy3 个月前

Here I was thinking MOBA games would be the first to have a world model

Eddy Xu 的头像
Eddy Xu3 个月前

woahh

Mark Grinev 的头像
Mark Grinev3 个月前

unless world models lead to net new creativity in gaming or a harry potter off the dome esk design why am i trying to play a knock version of an already existing game

Garry Fan 的头像
Garry Fan3 个月前

I'm being griefed by my teammate pls add a report button 9/10

Ronak Malde 的头像
Ronak Malde3 个月前

If AI progress stalled here I wouldn’t be too upset

justin 的头像
justin3 个月前

wtf

signüll 的头像
signüll3 个月前

this is incredibly cool.

Garry Fan 的头像
Garry Fan3 个月前

6 in the morning and I'm playing Rocket League in my browser holy shit this is cool!

Joseph Jojoe 的头像
Joseph Jojoe3 个月前

this is insanely cool

rohit 的头像
rohit3 个月前

the most impressive demo so far in 2026 - it's actually playable

Rahul Karajgikar 的头像
Rahul Karajgikar3 个月前

insanely cool stuff, as a rl player this is very exciting

josh halliday 的头像
josh halliday3 个月前

i am going to come and destroy MIRA at rocket league 🤣🤣

James Xu 的头像
James Xu3 个月前

this is insane 🚀 super cool work

thom 的头像
thom3 个月前

God we are taking off so hard

offmylawn 的头像
offmylawn3 个月前

what the actual

Chris Klein 的头像
Chris Klein3 个月前

Multiplayer world models before GTA 6.

acai 的头像
acai3 个月前

@0xm1kr @doppelfun

Alvaro L 的头像
Alvaro L3 个月前

Fascinating

Ethan Goodhart 的头像
Ethan Goodhart3 个月前

insane

Seth  Phillips 的头像
Seth Phillips3 个月前

Our friends just played it, absolutely incredible work.

venivci 的头像
venivci3 个月前

oh.

AL Robin 的头像
AL Robin3 个月前

The name is the same as the AI agents inside Telegram by The Open Platform :)

yang 的头像
yang3 个月前

Curious where medal data was used in this process since it says it was trained off bot data exclusively. “We train on roughly 10,000 match-hours of recorded 2v2 Rocket League matches, generated entirely by self-play between game-playing bots; no human gameplay at all. “

Team Reagent 的头像
Team Reagent3 个月前

@FelionSpike Wanna try it out? I absolutely called the fact that we would replace entire games with world models with a persistence mechanism

Rey Neill 的头像
Rey Neill3 个月前

Latency is rough, but impressive work!

Griz 的头像
Griz3 个月前

watching a model dream the next frame straight from your inputs with no engine underneath is the part that gets me

Neheart 的头像
Neheart3 个月前

10k hours just from bot-generated data is wild, curious how it handles real player unpredictability versus what the bots taught it.

Apurva Mishra 的头像
Apurva Mishra3 个月前

This is sooooooo coooool!!!!!

Sebastian Buzdugan 的头像
Sebastian Buzdugan3 个月前

20 fps matters, but recovery from bad player inputs decides if it feels playable

相关视频

At Avalon we are building "Real-time creating" - the ability to generate gameplay ready persistent worlds prompted from text. While others are building real-time video world models, Avalon is building real-time world generation inside a fully playable, persistent multiplayer engine. Internally running at 3840×2180 at 60 FPS. Built on Unreal Engine. Multiplayer by default. Persistent by default. Gameplay-ready by default. This is not a video latent replay. Not a simulation of interaction. It is a real 3D world with physics, logic, and authoritative multiplayer state. Avalon is trained on proprietary Avalon interaction data and powered by a hybrid system that combines language understanding, 3D model generation, procedural systems, and structured gameplay logic synthesis. Players can walk through a live world and generate environments, assets, mechanics, and entirely new gameplay modes using natural language. We accomplish this through a combination of 3D model generation, game logic generation based on our proprietary systems, and AI driven world creation. While other players are inside it. Changes persist instantly. State is synchronized in real time. Creation happens inside the world, not outside of it. Describe a biome. Spawn a civilization. Create a survival mode. Build a dungeon crawler. Launch a new game inside the world. Avalon interprets intent and integrates it directly into the live multiplayer environment. This is not a world model predicting video. This is a gameplay engine that understands language. If you can describe it, you can build it. And others can walk into it instantly.

AVALON

66,037 次观看 • 7 个月前