GoKawiil
Tech News
Cloudflare has released two open-weight models, Clef and Clef-flash, built for structured decision tasks like yes/no answers, multiple-choice, and ranking. The models use frozen, post-trained versions of Qwen backbones, support images, video, and a 64K context window, and are downloadable via Hugging Face.
Georg Zoeller describes using an AI model, Qwen-3.8-Flash-Next, to reverse engineer the 1989 game War of the Lance and port it to a Three.js browser engine over about a week, limited by running a single lane due to VRAM constraints. He is now further modernising the port into a Total War/Heroes of Might and Magic-style game while retaining the original game's underlying logic, and separately tested Fable 5.1 for generating Blender models via Python scripts, producing roughly 80-100 usable models per week of subscription.
A commentary piece argues that DHH's Omarchy Linux distro is being built and marketed around heavy use of large language models and agentic AI workflows, following a lengthy podcast interview in which DHH discussed his use of AI tools. The piece also notes Alibaba's Phoronix-reported Qwen Book announcement, which frames plans for Omarchy to become the operating system for Alibaba's proposed agentic computer hardware, and mentions Michael Dell's donation and DHH's on-air praise of the Dell XPS.
Apple's AI research team has published LensVLM-9B, a 9-billion-parameter vision language model that processes long documents by first compressing text into image form and then selectively decompressing only the pages it judges relevant, using learned tools. The model, code and demo scripts have been released on GitHub and Hugging Face under Apple's Machine Learning Research Model License and Sample Code License, with configurable compression ratios of 5x, 10x, or 15x.
Alibaba Cloud has launched Qwen Image 2.1, an open-weight image generation model with only 7 billion parameters, small enough to run on older consumer GPUs like the RTX 3090. The model adds native transparent-background generation and can combine up to 10 reference images into a single consistent output, and Alibaba's own benchmarks claim it outperforms closed models including Google's Nano Banana 2.0. Unlike its predecessor, the new licensing terms bar commercial resale without a separate agreement from Alibaba.
At Apsara Conference 2026, Alibaba CEO Eddie Wu introduced the Zhenwu V900 accelerator, claiming it delivers three times the performance of the prior M890 chip and calling it China's most powerful AI chip. The V900 offers 216GB of GPU memory, 1,200GB/s inter-chip bandwidth, can scale to clusters of up to 500,000 cards, and is set for mass production in Q1 2027, earlier than the Q3 2027 date Alibaba previously targeted.
An AI researcher briefed members of Congress and staff on the current state of open-weight versus open-source language models, framed within U.S.-China competition. The briefing distinguished closed API-only models like GPT-4 and Claude from open-weight models such as Meta's Llama, Alibaba's Qwen and DeepSeek, and from fully open-source models like the Allen Institute's Olmo, noting Chinese firms have led in open-weight releases since around April 2025.
US officials removed a Chinese-made Qwen AI search tool from the Federal Register website on Wednesday after social media users spotted it powering search of public regulatory comments. The tool, made by Alibaba, had been live for at least a day before removal, and neither the National Archives, the White House, nor the FBI has explained how or why it was deployed.
Following a rushed 'Lite' GGUF release four days after Qwen 3.8 27B launched, the team has now finished fully optimized ShapeLearn quantizations and benchmarked them against both the Lite versions and rival quants. The full models push the quality-versus-speed tradeoff further, with all five new variants topping the performance frontier across six GPU test configurations.
Tom's Hardware Premium's weekly roundup details extensive benchmarking of the Qwen 3.8 27B language model on consumer-accessible hardware, including an RTX 5090, Mac Mini, DGX Spark and Strix Halo systems, showing it can approach top-tier AI performance without cloud API costs. The issue also covers IFA show reports noting a market split between ultra-light MacBook-style laptops and pricey agentic AI PCs, squeezing out affordable mid-range machines, plus a note on Ajinomoto's role in supplying materials for the chip industry.
Tom's Hardware benchmarked Alibaba's newly released Qwen3.8 27B open-weight AI model across discrete GPUs like the RTX 5090, RTX 4090, RTX 3090, and AMD/Intel cards, as well as unified-memory systems including the DGX Spark, Mac Studio, and Ryzen AI Halo. Despite the four-bit quantized model needing only about 17GB of VRAM, the tests found that software stacks and inference engines often bottleneck real-world performance, affecting time-to-first-token and throughput once context windows fill up.
A developer ran the same prompt—building a self-contained sci-fi hangar scene with Three.js, including hovering drones, animated lights, and camera paths—across 10 combinations of AI models (including GLM, Luna, SOL, Astra, and Qwen variants) and coding harnesses like Codex, OMP, OpenCode, and DSH. The test tracked metrics such as completion time, token usage, tool calls, error rates, and whether the model verified its own output by opening the file in a browser and checking screenshots.