#开源生态

Open-sourcing GenOffice, the world's first full-featured open-source AI Office for PC and Mac.  GenOffice is free for everyone, ad-free. All the editing tools you'd expect, no strings attached. How it started? One engineer, one week, $10,000 in tokens. That's what it took to build the GenOffice Alpha. Help shape the AI office experience We invite you to build the rest of it with us. Join the GenOffice group chat on GenTeam and bring your requests straight to the agent. Not by writing code, but by telling it what you want. You say what you need, and every piece of your feedback becomes part of the product. What's in GenOffice? Docs, Sheets, Slides, and PDF, with all the editing tools you already know. And with Genspark Super Agent built in, powered by top-tier models, it digs into research, crunches your data, then writes the doc or builds the deck for you, using your Genspark credits. This is an Alpha, and we mean it. Come tell us what's broken and what's missing. Everyone who jumps in with real feedback gets 1,000+ Genspark credits as a thank you. GitHub 👉 https://github.com/genspark-ai/genoffice Download it now 👉 https://www.genspark.ai/genoffice

We put explanations in the license that there are no regional restrictions. However, in the EU we follow conventions similar to LTX and Hunyuan; while in the US, due to local laws and regulations and the ongoing legal situation with Disney, users are required to complete an additional form and sign a waiver.

Qwen3.8 Max by @Alibaba_Qwen is live on OpenRouter. The new flagship has 2.4T parameters (95B active) and is built for long-horizon coding, research, and multimodal agent work. Open weights are due next week, a first for a Qwen Max-class model. https://openrouter.ai/qwen/qwen3.8-max

Generate Videos with Minimax H3 locally on your Mac! Phosphene 3.4.0 from @AIBizarrothe just added support for @MiniMax_AI H3, powered by MLX. That was fast! I am playing with it myself right now and it's very usable!

Today, we’re releasing Athena (mvrko-sim-1), the flagship Large Event Model from @markopolo_ai that beats GPT-5.6, Claude Opus 4.8, and Claude Sonnet 5 in OPeRa public benchmark for Shopping behaviour prediction! Athena introduces a completely different approach to understanding digital behavior. Most systems record actions after they happen. Athena models the sequence behind them to predict what a shopper will do next and exactly where they will do it, before the action even takes place. Athena vs. frontier models: We evaluated Athena on the full OPeRA test set alongside GPT-5.6, GPT-4.1, Claude Sonnet 5, and Claude Opus 4.8. Athena achieved 24.5% strict exact match scoring the highest among the five models evaluated. Athena is not a general-purpose language model. It is a Large Event Model built to understand behavior as a connected sequence. It combines the page someone is viewing, the actions that brought them there, and the goal they are trying to complete and then predicts the exact next browser action and target element. That distinction matters. Today’s digital systems largely react after someone clicks, abandons, purchases, or leaves. Athena creates the foundation for systems that can understand what is likely to happen before the outcome. Together with @RubaiyatFarhan, and the team at @markopolo_ai, we built Athena from years of work on shopper behavior, digital journeys, and behavioral intelligence. Commerce was our first proving ground. Great thing is, we’re now releasing Athena publicly so builders can adapt the model to new datasets, industries, and problems. We’re seeing early usecases in Cyber Security, Gaming, Mobile App Ecosystem, Retail and more! Athena is now live on Hugging Face! We built the foundation for relevance and predictability! Now, let’s see what the world builds on top of it. Links in the first reply.

H3 is the first model to reach sd-level quality and, most importantly, it’s open weights. Personally I think this is one of the biggest steps forward for both the ecosystem and the tech as a whole. From today on, we won’t be the only ones here for sure.

MiniMax H3 is native in ComfyUI, day zero with the weights. Model highlights: → Text-to-video, prompt only → Image-to-video → First-and-last-frame - control the opening frame, the closing frame, or both → Reference-to-video - carry a subject, a motion, or a voice through the clip → In-place editing - modify a shot you already have instead of regenerating it The capability @MiniMax_AI leads with is multimodal context understanding, and it's what collapses those five tasks into one model. H3 takes images, audio, and video together and resolves them against a prompt that explains how they relate. Download the workflows & learn more in our blog 👇

MiniMax H3 open weights are live on fal, day 0.

I’m seeing way better videos coming from H3 than 2.5 on the timeline. And more of them. I think cost is a big part of it, plus the speed of inference, and custom audio uploads working as intended. My thesis holds still that at a certain point unless you are actually fixing the pressure points creatives have with your tools, cost and speed are going to win. Unlimited generations are NOT a pipe team. It all comes down to hourly costs divided by users. At some point if you can scale you’ll run at a surplus and can run unlimited.