# HY-Himmel Technical Report: Hierarchical Interleaved Multi-stream Motion Encoding for Long Video Understanding

arXiv: https://arxiv.org/abs/2605.08158 · HTML: https://mlsanity.com/paper/2605.08158/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Video understanding with MLLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 30 of its 48 references that are in the map.
- First posted on arXiv 4 May 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Video understanding with MLLMs](https://mlsanity.com/mllm/video-understanding-with-mllms/)

## Builds on (most cited first, 30 of 30)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [RAFT: Recurrent All-Pairs Field Transforms for Optical Flow](https://mlsanity.com/paper/2003.12039/) (2020)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [MVBench: A Comprehensive Multi-modal Video Understanding Benchmark](https://mlsanity.com/paper/2311.17005/) (2023)
- [LLaVA-Video: Video Instruction Tuning With Synthetic Data](https://mlsanity.com/paper/2410.02713/) (2024)
- [LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models](https://mlsanity.com/paper/2311.17043/) (2023)
- [LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding](https://mlsanity.com/paper/2410.17434/) (2024)
- [Perception Test: A Diagnostic Benchmark for Multimodal Video Models](https://mlsanity.com/paper/2305.13786/) (2023)
- [LongVILA: Scaling Long-Context Visual Language Models for Long Videos](https://mlsanity.com/paper/2408.10188/) (2024)
- [Flash-VStream: Memory-Based Real-Time Understanding for Long Video Streams](https://mlsanity.com/paper/2406.08085/) (2024)
- [Adaptive Keyframe Sampling for Long Video Understanding](https://mlsanity.com/paper/2502.21271/) (2025)
- [Cambrian-S: Towards Spatial Supersensing in Video](https://mlsanity.com/paper/2511.04670/) (2025)
- [Compressed Video Action Recognition](https://mlsanity.com/paper/1712.00636/) (2017)
- [TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning](https://mlsanity.com/paper/2410.19702/) (2024)
- [Video-LaVIT: Unified Video-Language Pre-training with Decoupled Visual-Motional Tokenization](https://mlsanity.com/paper/2402.03161/) (2024)
- [LOOK-M: Look-Once Optimization in KV Cache for Efficient Multimodal Long-Context Inference](https://mlsanity.com/paper/2406.18139/) (2024)
- [BOLT: Boost Large Vision-Language Model Without Training for Long-form Video Understanding](https://mlsanity.com/paper/2503.21483/) (2025)
- [HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding](https://mlsanity.com/paper/2601.14724/) (2026)
- [Aligning Multimodal LLM with Human Preference: A Survey](https://mlsanity.com/paper/2503.14504/) (2025)
- [CoPE-VideoLM: Leveraging Codec Primitives For Efficient Video Language Modeling](https://mlsanity.com/paper/2602.13191/) (2026)
- [ReMoRa: Multimodal Large Language Model based on Refined Motion Representation for Long-Video Understanding](https://mlsanity.com/paper/2602.16412/) (2026)
- [CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference](https://mlsanity.com/paper/2604.06036/) (2026)
- [ShotFinder: Imagination-Driven Open-Domain Video Shot Retrieval via Web Search](https://mlsanity.com/paper/2601.23232/) (2026)
- [Beyond Closed-Pool Video Retrieval: A Benchmark and Agent Framework for Real-World Video Search and Moment Localization](https://mlsanity.com/paper/2602.10159/) (2026)

## Abstract (arXiv)

Long video understanding with multimodal language models suffers from three compounding bottlenecks: heavy decode cost to obtain dense RGB frames, token growth with frame count, and weak motion perception under sparse keyframe sampling. Existing remedies either prune visual tokens after decoding, which leaves the expensive RGB pipeline untouched, or they discretise codec motion vectors with a generative tokeniser, which couples motion modelling to a heavy pre-training stage. In this paper we present HIMMEL, a hierarchical video-language framework that allocates semantic and motion capacity along separate paths. A small set of sparse semantic anchor frames is routed to the expensive host ViT to ground object identity and scene layout, while the far denser inter-frame intervals are encoded by a lightweight compressed-domain tri-stream adapter that distils motion evidence from motion-vector maps, residual maps, and an I-frame context branch into aligned motion tokens. These tokens are injected into the LLM via a differentiable placeholder mechanism, after a dedicated Stage 1 contrastive alignment that places the motion representation in a geometry compatible with the frozen visual backbone. We further show that an InfoNCE alignment objective beats MSE regression by +1.5 pp. On Video-MME, HIMMEL surpasses the dense 32-frame baseline by +2.3 pp (61.2 to 63.5%) while using 3.6 times fewer context tokens and running end-to-end in 1.03 s per question instead of 2.75 s. Switching to a stronger Qwen3-VL-8B host pushes HIMMEL to 64.9%, narrowing the gap to much larger proprietary systems while keeping inference on a single NVIDIA H20 GPU. Extensive ablations across stream composition, motion-encoder family, fusion mode, alignment objective, anchor count, LoRA rank, and video duration confirm that the full tri-stream contributes to the observed gains, and that the benefit grows with video length.