AI Briefing#qwen
The latest AI and developer news about #qwen, with the original source and a short summary.
The latest AI and developer news about #qwen, with the original source and a short summary.
A lightweight end-to-end document structuring OCR model based on Qwen3.5-0.8B has been released.
Unsloth has released NVFP4 quantized models for Qwen3.6 that boost inference speed by up to 2.5x on Blackwell GPUs.
China pledged to expand the AI ecosystem through open-source models like DeepSeek and Qwen at the UN.
It explains the hardware setup and Docker configuration needed to run SOTA LLMs and STT models in a local environment.
Distillation, which began as a model compression technique, has now become a core technique for transferring reasoning ability between models and the center of a copyright controversy.
The Qwen team announced HydraHead, an architecture that combines FA and LA at the attention-head level to improve long-context processing performance.
Qwen Image Agent has dramatically improved text-to-image generation performance through planning, reasoning, and search capabilities.
A large-scale mutual grading of 55 LLMs confirmed significant evaluation bias according to model family.
The Qwen team has released Qwen-AgentWorld, a native language world model that simulates 7 domains.
The Qwen team has unveiled three robot foundation models that connect vision-language understanding to physical action.
LangSmith partnered with Fireworks to fine-tune a Qwen model, developing a 'Perceived Error' detection model that cuts costs by 100x.
Allegations have surfaced that the Rio city government's supposedly self-developed model is actually the result of weight merging of existing models.
The Rio de Janeiro city government released a 397B-scale open source model fine-tuned from Qwen.
Video input capability has been added to llama.cpp, enabling video analysis with Gemma and Qwen models.
Domino, a technique that separates modeling from drafting to maximize the efficiency of Speculative Decoding, has been unveiled.
An analysis alleges that Alibaba T-Head's asllm engine uses the open-source sglang without authorization and has left a serious bug unaddressed.
Introducing On-policy distillation (OPD), a core training technique used in the latest models such as Qwen and DeepSeek.
InstinctRazor technology has been unveiled, running a 122B MoE model even in an 8GB VRAM environment by placing experts on the CPU.
A PR applying post-norm hidden states to improve Qwen model MTP performance has been submitted to llama.cpp.
At Microsoft Build, a technical session on scaling agentic AI from on-device to cloud will be presented.
Qwen3.7-Plus, a multimodal agent that unifies vision and language to perform GUI and CLI tasks, has been unveiled.
Alibaba's Qwen team officially released a new model, Qwen 3.7 Plus.
A Qwen3.6 GGUF model has been released that integrates an MTP head, enabling self-speculative decoding without a separate draft model.
NVIDIA has released a quantized version of Alibaba's Qwen3.6-35B model using the NVFP4 data type.