Distribution Notices & Licensing

1. Core Inference Engine

The MasiDyn native local completion engine (masidyn-completion.exe) is a custom build built on top of the open-source llama.cpp project. We attribute the original work to Georgi Gerganov and all contributors of the llama.cpp project.

License: MIT License

MIT License

Copyright (c) 2023-2026 Georgi Gerganov

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

2. Microsoft BitNet b1.58 Models

The natively trained 1-bit / ternary model weights used for ultra-fast CPU inference (including bitnet-b1.58-2b-4t and bitnet-b1.58-3b) are based on the BitNet research architecture developed and open-sourced by Microsoft Corporation.

License: MIT License

Copyright (c) Microsoft Corporation. All rights reserved.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

3. Falcon 3 1.58-Bit Models

The Falcon3 1-bit models (including falcon3-1b-instruct-1.58bit, falcon3-3b-instruct-1.58bit, and falcon3-7b-instruct-1.58bit) are developed by the Technology Innovation Institute (TII).

License: TII Falcon License 2.0

These models are distributed under the TII Falcon License 2.0 (which is constructed on top of the Apache License 2.0). Any distribution of the models or their derivatives must retain a copy of the license and include clear attribution to the Technology Innovation Institute.

4. Llama 3 1.58-Bit Model

The Llama 3 1-bit model (llama3-8b-1.58-100b) is a ternary derivative quantized version of the Meta Llama 3 8B model.

Attribution Notice: Built with Meta Llama 3

This model is distributed under the Meta Llama 3 Community License Agreement. Any distribution or hosting of this model must carry the prominent attribution "Built with Meta Llama 3" and include a copy of the Meta license agreement.

5. MasiDyn Standard Scale Models

The standard scale local GGUF models configured in the MasiDyn local registry are attributed to their respective authors:

• Qwen 2.5 & Qwen 3 (0.5B, 0.6B, 14B): Developed by the Alibaba Qwen Team. Licensed under the Apache License 2.0.
• Gemma 2 (2B, 9B): Developed by Google LLC. Subject to the Gemma Terms of Use.
• nomic-embed-text: Developed by Nomic AI. Licensed under the Apache License 2.0.