ModelsSelected10月14日 23:23

VideoJev open-sources 4B and 35B video understanding models

Aran Komatsuzaki (论文推介)@arankomatsuzaki

Post

Why we picked it

Trading deliberate reasoning for a fast single-pass design makes video QA over 100x quicker, which matters for latency-sensitive applications like live video analysis. Builders working on video agents or segmentation can grab the 4B or 35B checkpoints and benchmark against their existing backbones using the reported mIoU numbers.

VideoJev is a System One model for video understanding and interaction, now open-sourced in 4B and 35B variants. The model delivers a claimed >100x speedup on video QA compared with thinking-based approaches. It also posts +6.7 and +10.8 point mIoU gains over 4B and MoE backbones respectively.

Original · Aran Komatsuzaki (论文推介)

VideoJev: System One Model for Video Understanding and Interaction

- >100x QA speedup vs. Thinking - +6.7 / +10.8 pts mIoU gain over 4B / MoE backbones - Opensourced 4B and 35B variants https://t.co/Ja33rRvxhI

#VideoJev #video generation #open-source models #multimodal

Keep reading

Recent stories with the same tag or section

Models Builders17:23

Chinese content米哈游联合清华、北大等高校开源 GameScaling 游戏模型，覆盖 0.8B 至 27B

米哈游和清华北大把游戏操作模型开源了，能自己玩《我的世界》和《毁灭战士》，最小 0.8B 单卡就能跑，可以拿来折腾。

#GameScaling #米哈游 #Qwen3.5 #开源模型

IT之家Source

Models BuildersSelected17:09

Tencent Hunyuan releases WorkForge agent training framework

Training agents on verifiable, evidence-grounded tasks addresses the core bottleneck of RL for long-horizon work: reward reliability. The 16.7K-environment scale makes this one of the larger open efforts in agent environment construction. Engineers building work or office agents can study the dual-verifier design for their own training pipelines.

#Tencent Hunyuan #WorkForge #agents #AI agents

官方账号腾讯混元@TencentHunyuanSource

Models Builders16:38

Chinese content中电信星辰实验室开源意图理解模型 Xing4.0-29B-Intent

中电信星辰实验室开源了 29B 的意图理解专用模型 Xing4.0-29B-Intent，做对话系统和意图识别的可以拿来试试。

#Xing4.0-29B-Intent #星辰 #中电信 #开源模型

量子位@量子位的朋友们Source

Models BuildersMulti-sourceSelected16:36

Muse Glimmer 30B: Meta's distilled agent model runs on one consumer GPU

Distilling from a large teacher into a small model specialised for long tool-use chains may matter more than raw size for local agents, a trade-off the OSWorld gap makes visible. Builders running agents on local hardware can use it as a drop-in for multi-tool MCP workflows without cloud dependency. Compared with Qwen3.6-27B, it wins on tool orchestration but loses on general OS-level tasks.

#Muse Glimmer #Meta #open-source models #agents

Event