#开源生态

muse spark open. why 2026 won't be like 1984 

1/ muse spark open is the most capable US open-weight model to date. 816B total params, 42B active. 1M-token context. natively multimodal. built for coding and agentic work like long-horizon tasks, hard software engineering, multi-step reasoning. america! 🇺🇲

muse spark open. why 2026 won't be like 1984 

For 30 years, technology has moved in one direction: turning information into intuitive interfaces Somehow, the smartest software we’ve ever built took us back to a textbox We’ve spent the last 2 years at @openuidev trying to fix that. 6 months ago we open-sourced it: 10K+ GitHub stars, 6M+ downloads ChatGPT shipped Intelligent UI and showed everyone what it looks like Today we’re releasing OpenUI 1.0 bringing it to every other agent

Meet NVIDIA Nemotron 3.5 Super VL Our new open model combines reasoning and tool use with text, image, document, and video understanding for long-running AI agents. 120B total parameters with 12B active 🧵

Chat models return messages. Decision models return decisions. Workflow models execute workflows. TypeLLM introduces the first workflow model. Run an entire workflow in a single API call. Simpler, faster, cheaper. Try it now: https://typellm.ai/examples/ticket-workflow

Ling 3.1 Flash running locally beat GPT-6 Luna at LEGO Eiffel Tower 🧱 @AntLingAGI just dropped Ling 3.1 Flash open weights and it crushed GPT-6 Luna on the same @threejs prompt for 2.8x cheaper Run AI models locally -> http://atomic.chat

MiniCPM-V 4.7 is now open source on ModelScope: a 35B-A3B MoE multimodal model built to run on-device without compromising performance. Apache 2.0. 🤖 https://modelscope.ai/collections/OpenBMB/MiniCPM-V-47 📱 First open-source MoE with smooth, lossless on-device inference, running at 13+ tokens/s on iPhone 17 Pro Max. 🔥 New Pareto frontier: in Instruct mode, it reaches a higher overall score in less inference time than Qwen3.6 35B-A3B. ⚡ Mixed 4x/16x visual token compression lets developers balance visual detail and inference cost per task. 🛠️ Two sizes (35B-A3B and 1B) with FP8, GPTQ, AWQ, BNB, and GGUF variants. Works with SGLang, vLLM, llama.cpp, Ollama, SWIFT, and LLaMA-Factory. 💻 Runs on iOS, Android, and HarmonyOS, with day-0 support for Rockchip, Houmo, Intel, AMD, and Arm.

We are releasing Ultramarine 1 Flash, a frontier-class model built for knowledge work. It is open-source and achieves similar performance to Fable 5, while being 60x cheaper. It was post-trained by a team of 5 for ~$300k all-in. Available on API now, with weights releasing in 2 weeks.

Nathan Lambert, former post-training co-lead at the Allen Institute for AI, now launching Trillium, a nonprofit open AI lab, joins a16z's Lisha Li to discuss whether open models can really keep up with the frontier. When he tells researchers in China that an average OpenAI researcher reportedly has five thousand H100s to work with, "their jaw just falls to the floor." In his view, Chinese labs have the talent but are mostly short on compute. He also thinks the distillation story is overblown. Based on the research, he puts its edge at a few months, and says "there's a bit of cope" in thinking only American labs can build good models. A lot of frontier progress outside math and code comes down to labs buying 1) data and 2) RL environments. It's a multibillion-dollar market, and some of it is sloppy enough to teach models to reward hack. That's a big part of why Trillium exists. Nathan wants to give away the environments other people would sell, and show exactly how his team uses AI to build AI, including where the models still get stuck. His bet is that the more people can see how this works, the better prepared everyone is for what's coming. 0:55 Can open models keep up with the frontier? 2:05 What it takes to reach the frontier 3:55 Open weights and the fast-follower playbook 8:25 Chinese labs: equal talent, less compute 9:55 Why distillation is harder than it looks 14:55 Is distillation theft or competition? 16:20 The lab flywheel vs owning your weights 21:45 Can general models beat every specialist? 26:30 AI at 1% of today's cost in 4 years 27:15 The RL env boom and reward hacking 31:30 What Trillium will build in the open 35:30 Breaking into AI in the age of AI slop 38:05 Open-sourcing how AI builds AI 44:25 Why Nathan isn't pre-training 49:45 Agent swarms and why lab secrets leak 54:25 Fast takeoff vs risks already here YouTube: https://youtu.be/BKuK0M6oDTI @natolambert @trillium_labs @lishali88

Models cost millions to train and minutes to copy. Fine-tuning and distillation can erase where they came from. Introducing the Fingerprinting Leaderboard, an open benchmark for model provenance — built so the proof lives in the weights, not the paperwork. https://fingerprint.sentient.foundation

🎬 35B MoE multimodal AI. On your phone. With just 3GB of memory. 🤯 This is MiniCPM-V 4.7 in action: reading real-world signs, planning around constraints, parsing complex documents, and reasoning over visual data. Powered by LES (Lookahead Expert Scheduling), the 35B-A3B model brings efficient MoE inference to mobile devices, while the lightweight 1B variant delivers capable multimodal understanding on resource-constrained hardware. 🛠️ With Self-Replay Fine-Tuning (SRFT), developers can customize models for new tasks while maintaining general performance. From iOS to Android, MiniCPM-V 4.7 makes powerful multimodal AI more practical to run locally. Try it 👇 🔗 http://github.com/OpenBMB/MiniCPM-V 🤗 http://huggingface.co/collections/openbmb/minicpm-v-47 Self-Replay Fine-Tuning: http://openbmb.cn/en/blog/self-replay-fine-tuning