Alibaba Releases R1-Omni: First Full-Modality Emotion Recognition with DeepSeek-Style RLVR

Discover R1-Omni: Alibaba's open-source full-modality LLM that integrates DeepSeek-style RLVR for enhanced emotion recognition across video, audio, and visuals.

"AI Disruption" Publication 5000 Subscriptions 20% Discount Offer Link.

---

For the first time, DeepSeek’s RLVR is applied to a fully modal LLM—that is, one that includes video!

In the blink of an eye, the Ali Tongyi Laboratory’s Bo Liefeng team has rolled out another open-source release—R1-Omni is here.

And once again, it’s in Hangzhou. What exactly are they up to?