Skip to content

Formats

Model releases latest news

Models for games: updates and open releases of world models, 3D and animation generation models, and base models for game agents.

39 picks21 in the last 30 days76 collected in all

Updated

Picks archive · Page 2

Oct 28Wed21–39
  1. Google DeepMind53

    DeepMind unveils next-generation world simulation model Genie 4

    Google DeepMind has unveiled Genie 4, a next-generation world model that generates interactive simulated worlds in real time, with native audio and conversational interaction. Users can change the weather or camera angle, or add characters and objects, using natural language, and Gemini adjusts prompts to keep the scene coherent. It is currently available as a research preview to trusted testers. Google AI Ultra subscribers and early Gemini API partners will gain access in the coming weeks, though visual detail and object persistence may still drift over time.

    Why it matters: It puts native audio, real-time steering, and Gemini reasoning into a single world simulation, offering a look at viable paths for interactive world models.

Oct 22Thu
  1. fal36

    HeyGen Video and HeyGen Voice now live on fal

    fal adds HeyGen Video and HeyGen Voice models. Video clips include built-in dialogue, ambient sound, and sound effects, with text-to-video, image-to-video, and reference-image video generation at 5 to 15 seconds per clip and up to 2K. Voice supports 39 languages and instant voice cloning, with a faster Turbo tier.

    Why it matters: The clip length, resolution, and language count given in the announcement are fairly specific, making it easy to compare with other video and voice generation tools.

Oct 21Wed
  1. Tripo38

    Eden-Live now officially available on @reactorworld.

    Tripo's real-time generative rendering tool Eden-Live is now officially available on the Reactor platform, with a live online demo to try. According to a quoted post, it can repaint 3D scenes or prototypes into interactive worlds, change styles in real time using reference images and text, and allow prompts to be edited during runtime.

    Why it matters: Eden-Live offers an online demo on the Reactor platform, useful for developers interested in real-time visual repainting and scene stylization who want to evaluate integration options.

Oct 20Tue
  1. QbitAI53

    Meshy launches Meshy Edit, enabling local 3D model edits from a one-sentence description

    Meshy releases Meshy Edit, a local editing feature in which users box-select a region of a model and describe the change in one sentence, generating a new model in about a minute while the unselected parts stay unchanged. The feature is open to all registered users, costs 20 credits per edit, and supports high-precision models generated with Meshy 6 and Meshy 7.1. Every edit generates a new model and keeps the original; the new model can be edited further, downloaded, or sent straight to slicing software for 3D printing.

    Why it matters: It lays out the edit flow — box-select a region, describe the change in one sentence — and the price of 20 credits per edit, so 3D creators can judge whether it saves them the time of manually revising a model.

Oct 19Mon
  1. Viggle AI35

    Viggle AI releases JST-2, a physics-native world foundation model

    Viggle AI has released JST-2, a full-stack physics-native world foundation model. According to the announcement, its entire model stack is built in 3D rather than in pixel or video space. The company says it can generate 3D scenes within seconds, scale them in real time, and create characters with built-in motion using the same general architecture.

    Why it matters: Unlike most world models that learn in pixel or video space, it builds its entire model stack directly in 3D, making it a useful reference for comparing different technical approaches.

  2. QbitAI57

    VAST open-sources EdenLive real-time world model, over 24 FPS at 720P on a single consumer GPU

    VAST has open-sourced EdenLive, a real-time world model that generates frames live under any reference input and art style and keeps running continuously. According to VAST, on a single Nvidia RTX 5090 at 720P, its 5B model exceeds 24 FPS, with end-to-end latency in the hundreds of milliseconds. EdenLive handles only visual rendering, while world state and interaction rules are still maintained by the upstream interaction system. Model weights, inference code and a technical report are now available.

    Why it matters: The article explains the design split between real-time rendering and state maintenance, and gives speed and latency figures on consumer GPUs, helping readers judge its usable range in game prototyping.

Oct 16Fri
  1. ElevenLabs39

    ElevenLabs launches Instant Soundtracks to generate music and sound effects for videos in one pass

    ElevenLabs has launched Instant Soundtracks. After you upload a video, the model generates original background music and adds sound effects at suitable moments. You can also use the video's original audio as a reference so the new soundtrack stays close in style. The feature is now available on ElevenCreative, ElevenMusic, and ElevenAPI.

    Why it matters: The feature is aimed directly at video creators. It generates original music and sound effects in a single pass, and is worth comparing against existing video post-production workflows.

Oct 15Thu
Oct 14Wed
  1. fal34

    Mirelo SFX V2.0 is now live on fal

    Mirelo SFX V2.0 is now live on fal, generating sound effects from text prompts or video input. It can replace unwanted music and sounds while preserving dialogue, and produce coherent sound effects and ambient audio across video cuts.

    Why it matters: Beyond text prompts, it can generate sound effects from video input and replace unwanted music and sounds while keeping dialogue intact.

Oct 13Tue
  1. QbitAI38

    ByteDance world model debuts live at 2026 Wonderful Night, generating and freely exploring 3D scenes from a single sentence

    At the tech segment of Douyin Group's 2026 Wonderful Night, ByteDance showed its self-developed world model, which can generate a 3D scene from a single sentence and supports free exploration, real-time interaction, object generation and editing, and cross-scene connectivity. The model has not been officially released yet, and the live demo was presented by Tim of Yingshi Jufeng.

    Why it matters: Using ByteDance's world model live demo as an entry point, this compares the differing approaches of OpenAI, DeepMind and Nvidia to spatial modeling and physical AI training.

Oct 11Sun
  1. The Decoder65

    Odyssey opens world model Odyssey-3 in public research preview with a free online demo

    Odyssey released a public research preview of its world model Odyssey-3, letting users generate interactive environments from text prompts and explore them in real time, with developers able to apply for API access. The free online demo runs on Odyssey-3 Flash; the base model has 1.4 billion parameters and generates 832 × 480 video, while Odyssey-3 Pro supports 1280 × 720.

    Why it matters: On physics benchmarks, the original article gives both the official self-selected single-run score and the four-round average, so readers can judge the differences in statistical reporting for such world model results.

Oct 10Sat
  1. Tripo44

    Tripo P2.0 adds quad mesh generation, polycount variants, and local Mesh Edit

    Tripo P2.0 update supports cleaner quad mesh generation, polycount variants, and local Mesh Edit. Smart UV automatically handles seam placement and UV unwrapping, while Remove Lighting generates cleaner base colors for easier adaptation to different engines' lighting setups. Rigging offers a more complete humanoid skeleton and engine-ready presets, and Text to Motion has been added for fast animation generation, with the entry point at https://studio.tripo3d.ai.

    Why it matters: P2.0 puts quad mesh generation, polycount variants, and local Mesh Edit into a single workflow and provides engine-ready rigging presets, making it easy for artists to adjust assets to project needs.

Oct 9Fri
  1. Odyssey35

    Odyssey releases world model Odyssey-3

    Odyssey has officially announced the release of its world model Odyssey-3, calling it a big step toward having world models power physical AI and generate environments for training intelligences. The company says Odyssey-3 is now available to try and provides the link odyssey.systems/meet-odyssey-3. The tweet did not disclose model size, capability details, or availability scope.

  2. Odyssey35

    Odyssey releases foundation world model Odyssey-3

    Odyssey released the foundation world model Odyssey-3, saying it can support a range of applications including physical AI, human experiences and agent training. The company specifically notes that agents can learn from experience and accomplish goals inside Odyssey-3.

    Why it matters: Odyssey lays out how it positions Odyssey-3, so readers can see how foundation world models are being applied to agent experience learning.

Oct 8Thu
  1. Tripo36

    Tripo launches P2.0 Official in Tripo Studio

    Tripo's official account announced that P2.0 Official is now live in Tripo Studio, inviting users to try it out in its generation workspace. The entry point given in the post is https://studio.tripo3d.ai/workspace/generate?tab=smart_mesh&utm_source=twitter&utm_medium=social&utm_campaign=game-ready-pipeline&utm_content=p2-official.

Aug 31Mon
  1. Inworld AI Blog57

    Inworld AI releases Realtime TTS-2 realtime conversational voice model

    Inworld AI has released Realtime TTS-2, a new-generation realtime conversational voice model, now fully available in the Inworld API and the Inworld Realtime API.

    Why it matters: The original post covers multi-turn audio context, natural-language voice instructions and a first-audio latency of under 200ms, which readers can use to judge whether realtime voice conversation fits their own projects.

Jun 10Wed
  1. Inworld AI Blog40

    Inworld cuts prices across its voice and LLM stack, by more than half for most developers

    Inworld announced price cuts across its entire voice stack, with most developers seeing reductions of more than half at each layer of text-to-speech, speech recognition, LLMs and compute. The new prices are already in effect for all developers. Realtime TTS-2 drops to about $10/1M characters, speech recognition to about $0.10/hour, and Router can connect to 100+ models with no markup on third-party models.

    Why it matters: Inworld has lowered prices at every layer of its voice stack and published tiered per-unit rates for TTS, speech recognition and LLMs, which can be used to estimate inference costs for consumer-grade applications.

May 14Thu
  1. Inworld AI Blog49

    Inworld pairs Realtime TTS-2 with Stream Vision Agents for a realtime voice agent that can see and listen

    Inworld teamed up with Stream to build a reference implementation, Crashout Buddy, combining the latest speech model Realtime TTS-2 with Stream's open-source Vision Agents framework, so it can see facial expressions, hear speech and adjust its delivery in real time.

    Why it matters: The post shows how to turn visual signals and emotional context into speech delivery instructions, which readers can use to judge whether this realtime voice pipeline can be plugged into their own projects.