KlingAI's VideoNexus splits perception from reasoning for cheaper, more accurate long-video understanding
Kling-Keye · hf · 2026-11-18
KlingAI Research, with CUHK and HKUST, introduces VideoNexus, a training-free framework for long-video understanding that separates detailed perception from evidence-driven planning and reasoning.
- The planner decides where to inspect, what details to resolve, and how densely to sample; perception sub-agents process native audio-video independently and return compact, temporally grounded evidence.
- This avoids piling raw frames or per-frame descriptions into the planner's reasoning context, achieving higher accuracy at lower cost on long-video benchmarks.
More from Multimodal
- Former voice actor launches Voyager 1, an AI voice you can direct line by line — Scobleizer · 2026-11-18
- Recreating Whiplash's tense rehearsal room with MiniMax H3 and ComfyUI — hucun16 · 2026-11-18
- Claude Plus Unreal Engine 5.9 MCP Builds a Game Prototype in One Day — TomLikesRobots · 2026-11-18
- Suno's new Visuals feature pairs AI-generated music with AI-generated video — The Decoder · 2026-11-18
- Osmo claims human-level AI smell: 10M labels trained world's first olfaction foundation model — MIT Tech Review AI · 2026-11-18
- AURA 2.0: a streaming video LLM that speaks first, calls tools, and grounds visually — Shuhan Li · 2026-11-18