1
摩尔线程完成MiniMax H3多模态生成模型适配
国产GPU新突破：摩尔线程成功适配MiniMax多模态模型，AI算力生态再进一步
36氪获悉，今日，MiniMax正式开源多模态生成模型MiniMax H3。同日，摩尔线程基于AI训推一体智算卡MTT S5000及MUSA软件栈，率先完成H3的适配与运行。
2
3
MUNI: Multimodal Unified Latent Diffusion for Coherent Any-to-Any Generation
统一多模态潜在扩散模型实现任意模态间的连贯生成，无需配对数据，跨模态生成能力惊艳
arXiv:2606.16408v1 Announce Type: new Abstract: We introduce MUNI, an end-to-end multimodal latent diffusion framework for any-to-any generation that …
4
Show HN: RunAPI – one API for AI video, image, music, audio, and LLMs
一个API搞定10+AI服务，视频图像音乐音频LLM全囊括，开发者效率利器。
After building a few SaaS apps, I found that connecting many AI model APIs at the same time gets messy very quickly. So I built RunAPI, which lets you…
5
JAM-Flow: Joint Audio-Motion Synthesis with Flow Matching
联合人脸运动与语音生成，Flow Matching实现音视频同步合成，突破传统分离建模局限。
arXiv:2506.23552v2 Announce Type: replace Abstract: The intrinsic link between facial motion and speech is often overlooked in generative modeling, wh…