AI Briefing#vision-language-model
The latest AI and developer news about #vision-language-model, with the original source and a short summary.
The latest AI and developer news about #vision-language-model, with the original source and a short summary.
Cactus Compute's Mirage model uses 51.3M parameters to perform multi-task visual analysis on mobile NPUs, enabling text-only LLMs to answer visual questions without fine-tuning.
The 4B model uses 13.0% fewer model calls and 25.4% fewer output tokens than comparable systems after online RL.
Built on Qwen3-VL, WeVisOne-VL uses a 9.6M-parameter pixel decoder to unify object-level and pixel-level perception tasks.
The new 35B-A3B MoE model utilizes Lookahead Expert Scheduling to enable local vision-language processing on smartphones.
The new model achieves 13+ tokens/s decoding speed on iPhone 17 Pro Max using Lookahead Expert Scheduling.
The model runs entirely on a single 80 GB GPU without external services.
The project introduces webbrain-vl-2-450M, a specialized vision model trained on browser screenshots that runs locally via WebGPU.
The new 280M-parameter draft model accelerates LFM2.5-VL-3B inference by up to 3.13x on device and 2.66x on H100 GPUs.
Based on Qwen3.5-9B, LensVLM maintains original-level accuracy at a 4.3x compression rate and demonstrates performance up to 10.1x.
Alibaba DAMO Academy has open-sourced 'Damo Radar', a medical AI model that identifies cancer and over 150 diseases from CT scans.
inclusionAI has released Ling-3.0-flash-VL, a multimodal model featuring a 1M token context window and MoE architecture.
The Qwen team and Huazhong University of Science and Technology (HUST) have released Qwen-Drive-1.0, a vision-language model specialized for autonomous driving, achieving a score of 90.7 on the NAVSIM benchmark when applying RL.
NeoMME, a multilingual multimodal encoder that processes text and images with a single Transformer without a vision tower, has been released.
MirroS released the Code-as-World model and research results, which restore videos into simulator-executable code to enhance physical reasoning performance.
The Qwen team released Qwen-Drive-1.0, a vision-language foundation model dedicated to autonomous driving that integrates 3D perception and motion planning.
LG AI Research has unveiled an autonomous vision inspection framework that combines VLM and Agentic AI to eliminate human intervention from labeling to deployment.
The Sol model in OpenAI's GPT-5.6 lineup achieved overwhelming performance on vision benchmarks.
The Qwen 3.8 27B model has been released, featuring significant improvements in coding, research, agent tasks, and vision-language understanding.
Mistral has released OCR 4.1, featuring paragraph-level bounding box extraction and structural block labeling capabilities.
The next-generation open model Qwen3.8-27B, with enhanced coding and vision capabilities, has been released.
CohereLabs released a 2.4B vision-language model under the Apache 2.0 license.
Liquid AI released a 3.1B vision-language model with enhanced screen understanding and tool use.
CARE-X is presented as a chest X-ray VLM that combines generation and structured prediction.
VLX-Seek 1.5 10B has been released as an open-source model for real-world visual grounding.