Project ModelBay Launches to Protect Open-Source AI from Restrictions

A new community effort to keep open models available even if they get blocked.

Max Zanoga (@zanoga), the builder behind the well-known 32-GPU home AI datacenter, has launched a new project called ModelBay. The goal is straightforward: make sure open-source AI models stay available to everyone, even if governments or companies start to restrict them.

The timing is not an accident. In June 2026 the US government used export controls to pull Anthropic's Fable 5 and Mythos 5 models offline worldwide, and OpenAI released its GPT-5.6 model only to a small group of government-approved partners. Both moves followed a new executive order that set up government review of the most capable models before they ship. For now these controls apply to closed frontier models, but Max Zanoga argues that open weights could be next, and that the community should prepare before it happens rather than after.

ModelBay is a plan to keep open models reachable without depending on any single company, website, or country. It has four parts:

• Peer-to-peer distribution. Model weights are shared over BitTorrent instead of through a single host, the same way the first Llama and Mistral releases spread. A public list of file hashes lets anyone confirm they downloaded the real weights and not a tampered copy.
• Library and documentation backups. Running a model takes more than the weights. ModelBay keeps Git backups of the inference libraries, quantization tools, configs, and documentation needed to rebuild the full stack from scratch.
• A Tor mirror. A copy that is hard to take down, so the trackers and resources stay reachable even if the normal versions get blocked.
• Shared compute (planned). A future pool where people contribute a small slice of idle GPU time to run models for the community. Max Zanoga plans to start by sharing his own 32x RTX 3090 cluster when it sits idle.

Max Zanoga is well known in the local-LLM community for documenting his build from two GPUs to a 32-card cluster with 768GB of VRAM, running state-of-the-art open models locally on solar power. ModelBay comes out of that same hands-on work. He started thinking about peer-to-peer distribution after watching Hugging Face, where most people download their models, go offline more than once.

The weights-distribution piece is already under active development. Max Zanoga is now asking the open-source community for honest feedback on the idea, and wants to know how many people would be willing to keep and share hundreds of gigabytes of model weights to make the network work. The more people who take part, the stronger and harder to block the network becomes.

Smaug-72B Unleashed: Revolutionizing Open-Source AI on Hugging Face

Smaug-72B, the latest offering from Abacus AI, marks a significant milestone in the realm of open-source large language models (LLMs). Hosted on Hugging Face, an esteemed platform in the AI community for sharing and collaborating on AI models and tools, Smaug-72B has quickly ascended to the top of the LLM leaderboard, showcasing its exceptional capabilities in text generation and beyond. This model distinguishes itself by scoring an average of 80 across various benchmarks, a feat that sets a new standard in the field and highlights its superiority over other models like Mistral Medium in certain aspects.

The development of Smaug-72B is a testament to the innovative approaches applied by the Abacus AI team. Leveraging a fine-tuned variant derived from the Qwen-72B model, the team has employed specialized techniques aimed at enhancing reasoning and mathematical abilities, as evidenced by its impressive GSM8K scores. These techniques are not only pivotal for the model's current achievements but also signify a forward-looking strategy with plans to further refine and apply these methods to other high-caliber Mistral Models, including the miqu model, a 70B fine-tune of LLama-2.

The commitment to open-source innovation is further underscored by Abacus AI's plans to publish a research paper detailing the methodologies behind Smaug-72B. This initiative aligns with the broader goal of advancing the field of AI and making cutting-edge technologies accessible to a wider audience. The model's availability on Hugging Face allows researchers, developers, and enthusiasts to download, quantize, and experiment with Smaug-72B, fostering a collaborative environment where knowledge and resources are shared openly.

Furthermore, the collaboration between AWS and Hugging Face to enhance the accessibility and efficiency of generative AI applications underscores the growing ecosystem supporting these advanced models. AWS's infrastructure and tools, including Amazon SageMaker, AWS Trainium, and AWS Inferentia, provide a robust foundation for training, fine-tuning, and deploying models like Smaug-72B, ensuring that developers can optimize performance and reduce costs effectively.

In essence, Smaug-72B embodies the pinnacle of current AI research and development efforts, driven by a commitment to open-source principles and the pursuit of excellence in artificial general intelligence (AGI). As the AI community continues to explore and push the boundaries of what's possible, models like Smaug-72B serve as both a benchmark for future innovations and a resource for fostering creativity and problem-solving across various domains

The evolutionary tree of modern LLMs

 

Flourishing Canopy: The Era of Transformers

The canopy of our tree is dense with Transformer-based models, which have dominated the landscape since 2019. The GPT (Generative Pretrained Transformer) series, starting from GPT-1 and evolving rapidly to GPT-2 and GPT-3, has demonstrated a remarkable ability to generate human-like text. Meanwhile, models like T5 from Google and Microsoft’s Turing-NLG have pushed the boundaries of what's possible with language comprehension and generation.

Latest Blossoms: State-of-the-Art Models

Perched at the top of our tree are the latest and most advanced LLMs. Models like GPT-4 and others from various AI labs like OpenAI, Anthropic, and DeepMind are not only more powerful but also more nuanced in their understanding of language. They're capable of tasks ranging from writing essays to coding, and even creating art or music.

The Ecosystem: Open-Source vs. Closed-Source

An important aspect of our tree is the delineation between open-source and closed-source models. Open-source models, such as those from Hugging Face’s Transformers library, provide accessibility and transparency, allowing for widespread use and innovation. On the other hand, closed-source models, often developed by big tech companies, keep their inner workings under wraps, sometimes offering more powerful capabilities but less community insight and control.

Future Growth: The Path Ahead

As we look to the sky through the leaves of this tree, it’s clear that the future of LLMs holds even greater potential. With advancements in ethical AI, interpretability, and multi-modal capabilities, the next generation of LLMs is poised to be even more integrated into our digital lives.

The journey through the evolutionary tree of modern LLMs reveals a rapid and complex growth pattern, one that reflects both technological advancements and our deepening understanding of natural language. As we continue to nurture and develop these models, we can only imagine the heights they will reach and the ways they will transform our interaction with technology.

Exploring the Frontiers of AI Art: A Deep Dive into OpenDalle

In the rapidly evolving world of artificial intelligence, OpenDalle stands out as a beacon of innovation and accessibility in the field of AI-generated art. Born from the inspiration of OpenAI's groundbreaking DALL-E, OpenDalle is an open-source project that democratizes the power of AI in creating stunning visual art from textual descriptions.

The Genesis of OpenDalle

OpenDalle's journey began with the introduction of DALL-E by OpenAI, a state-of-the-art model known for generating highly detailed and contextually relevant images from textual prompts. While DALL-E impressed the tech world with its capabilities, it remained largely inaccessible to the general public due to its proprietary nature. This is where OpenDalle enters, aiming to bring similar technology to a wider audience.

Why OpenDalle Matters

Accessibility and Learning: OpenDalle offers a platform for enthusiasts, artists, and developers to explore the intersection of AI and art. Whether you're a student learning about AI or an artist experimenting with new mediums, OpenDalle provides the tools to delve into this exciting technology.

Community-Driven Innovation: Being open-source, OpenDalle thrives on community contributions. This collaborative approach accelerates improvements and leads to innovative uses beyond its original design.

Ethical and Transparent AI: OpenDalle encourages discussions about the ethical implications of AI in art. Its open-source nature allows for more scrutiny and understanding of AI's capabilities and limitations, fostering responsible use.

Applications and Possibilities

From generating unique artwork for personal projects to assisting in graphic design, OpenDalle's applications are vast. Educators can use it to demonstrate AI's capabilities in the classroom, while developers can integrate it into various applications, pushing the boundaries of what's possible in AI-assisted creativity.

The Future of OpenDalle

As AI continues to advance, projects like OpenDalle not only keep pace but also push the envelope. The future may see enhancements in image quality, diversity in style, and even more intuitive interactions between human prompts and AI-generated artwork.

Conclusion

OpenDalle is more than just a tool; it's a testament to the collaborative spirit of the AI community. It represents a step towards making advanced technologies more inclusive and accessible, and its impact on the world of AI and art is just beginning to unfold.

Unveiling the Future of AI Video: Introducing Stable Video Diffusion

Stability AI has announced the release of Stable Video Diffusion, a state-of-the-art generative AI video model, which is an advancement based on their image model, Stable Diffusion. This new model is adaptable for various video-related applications and is in a research preview phase, with the code and weights available on GitHub and Hugging Face. It is capable of generating videos with up to 25 frames at customizable frame rates. Although it shows promising performance, surpassing other models in user preference studies, it is currently intended for research purposes only and not for real-world or commercial use. Stability AI continues to expand its suite of AI models across different modalities, contributing to the field of AI with open-source solutions.

Unveiling MistralOrca: A Leap Towards Open-Source AI Excellence

In recent times, the realm of open-source AI has seen a monumental stride towards excellence with the advent of MistralOrca. This innovative model emerged from a collaboration that aimed to push the boundaries of what moderate consumer GPUs can achieve.

The narrative began with the meticulous creation of the OpenOrca dataset, a notable endeavor to replicate the dataset heralded in Microsoft Research's Orca Paper. The Alignment Lab team, known for their avant-garde approach, fine-tuned Mistral 7B using this dataset, employing OpenChat packing and Axolotl for training. The result? A release that boasts of GPT-4 augmented data finesse, a trait shared with its sibling model, OpenOrcaxOpenChat-Preview2-13B.

As of its release, MistralOrca carved its niche as the second-best model among those with a size less than 30B on the HF Leaderboard, outclassing all but one 13B model. This wasn't just a release; it was a statement of capability, a fully open model showcasing class-breaking performance, all while being accessible on moderate consumer GPUs. A tip of the hat to the Mistral team for pioneering this trail.

MistralOrca is not just a name; it's a testament to the seamless meld between robust technology and the open-source ethos. Want to give it a whirl? It's running on fast GPUs unquantized for an unparalleled user experience. You can find it [here](https://huggingface.co/spaces/Open-Orca/Mistral-7B-OpenOrca).

Beyond just a model, the team has provided a visual feast for data enthusiasts. Explore the full (pre-filtering) dataset on their Nomic Atlas Map, an endeavor to offer transparency and insights into the data that powers MistralOrca.

Now, for those who are on the lookout for quantized models, TheBloke has generously facilitated quantized versions of MistralOrca. Whether it's AWQ, GPTQ, or GGUF, a new horizon of exploration awaits [here](https://huggingface.co/TheBloke/Mistral-7B-OpenOrca-AWQ).

The tale doesn't end here; the ongoing journey sees the Alignment Lab in the throes of training more models, with the promise of exciting partnerships on the horizon. Stay tuned for sneak-peak announcements on their [Discord](https://AlignmentLab.ai), or delve deeper into the Axolotl trainer on the [OpenAccess AI Collective Discord](https://discord.gg/5y8STgB3P3).

The performance metrics of MistralOrca are nothing short of stellar. With a 105% performance of the base model on HF Leaderboard evaluations, it transcends the performance of all 7B models and all but one 13B model, averaging a score of 65.33. These evaluations were meticulously carried out using the Language Model Evaluation Harness, mirroring the HuggingFace LLM Leaderboard version.

The comparative analysis with the base Mistral-7B model unveils a 129% performance on AGI Eval, averaging 0.397, and a 119% performance on BigBench-Hard, averaging 0.416. These figures aren’t just digits; they are a testament to the colossal strides MistralOrca has made in the field.

The training regime was no less rigorous. Utilizing 8x A6000 GPUs for a marathon 62-hour training run, the team completed 4 epochs of full fine-tuning on their dataset. The commodity cost stood at a modest ~$400, a small price for a giant leap in open-source AI excellence.

In conclusion, MistralOrca isn’t merely a model; it’s a milestone in the open-source AI narrative. It epitomizes what collaborative effort, coupled with cutting-edge technology, can achieve. As the AI community waits with bated breath for what's next from Alignment Lab, MistralOrca stands as a beacon of what’s possible in the ever-evolving world of Artificial Intelligence.