PulseAugur
EN
LIVE 04:26:07
中文(ZH) 开源不只是「秀肌肉」

Sand.ai releases trillion-parameter open-source video MoE model

By PulseAugur Editorial · [2 sources] ·

Sand.ai has released MAGI-2-preview, an open-source, trillion-parameter Mixture-of-Experts (MoE) video generation model. This release provides researchers with a foundational infrastructure for studying large-scale MoE models and offers enterprises a new option for private deployment and fine-tuning, particularly in data-sensitive sectors. The move signifies a shift in the open-source vs. closed-source competition, with factors beyond raw generation quality, such as deployability and ecosystem support, becoming key differentiators. AI

IMPACT This release provides researchers with a foundational infrastructure for studying large-scale MoE models and offers enterprises a new option for private deployment and fine-tuning.

RANK_REASON The cluster reports on the release of a trillion-parameter Mixture-of-Experts video generation model by Sand.ai, which is a significant advancement in the field. [lever_c_demoted from frontier_release: ic=2 ai=1.0]

Read on 量子位 (QbitAI) →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

Sand.ai releases trillion-parameter open-source video MoE model

How we ranked this

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
The cluster reports on the release of a trillion-parameter Mixture-of-Experts video generation model by Sand.ai, which is a significant advancement in the field. [lever_c_demoted from frontier_rele…
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
54 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

Full methodology in our editorial standards.

COVERAGE [2]

  1. 量子位 (QbitAI) TIER_1 中文(ZH) · 鹭羽

    Open Source Is More Than Just "Showing Muscles"

    过去开源的视频模型大多都是百亿级，千亿参数直接改变的是整个行业的游戏规则。 对于 研究者 而言，这是第一次能够真正解剖千亿级视频MoE。 专家如何分工、路由是否稳定、通信怎样组织、Scaling规律能否复现，这些问题终于有了公开研究对象。 不必隔着论文空想，这对学术界和中小研究团队是实打实的基础设施升级。 对 企业 来说，开源意味着多了一个私有化部署和行业微调的选项。 尤其是金融、医疗、工业等数据敏感行业，数据能否留在本地、模型能否针对业务继续训练，重要性并不亚于一次生成效果。 再把视线放大至整个 行业 ，开闭源的竞争维度也将被重新定义。 以前是单纯效果

  2. X — MiniMax AI TIER_1 English(EN) · MiniMax_AI ·

    The power of open source! 🩵

    The power of open source! 🩵 Kudos to the local host community, it was released within a day and we’re already seeing MiniMax H3 running on Mac!

RELATED ENTITIES

RELATED TOPICS