What is MatrixLM?
MatrixLM is an independent AI lab in Brussels developing and training language models, model infrastructure, and experimental neural architectures.
FAQ
Questions about MatrixLM.
Learn more about our models, research, product access, billing, developer plans, and data handling.
MatrixLM
Does MatrixLM train its own models?
Yes. MatrixLM develops and trains its own language models rather than relying exclusively on third-party model APIs.
Where is MatrixLM based?
MatrixLM is developed in Brussels, Belgium, with the long-term ambition of contributing to a stronger Belgian and European AI ecosystem.
Is MatrixLM a single model?
No. MatrixLM is the project and platform. We work on multiple models and model approaches, including conventional language models and experimental architectures.
Models & research
What kinds of models does MatrixLM develop?
Our work includes conventional language models as well as experimental approaches to model architecture and computation. Different approaches are trained and evaluated independently.
Why experiment with new architectures?
Modern language models are dominated by a small number of architectural approaches. We think there is value in exploring alternatives while continuing to develop and evaluate established model architectures.
How do you evaluate experimental models?
Experimental systems are evaluated through controlled comparisons covering areas such as training behavior, retrieval, generalization, reasoning, and generation. This work is ongoing and current results should be treated as internal research rather than independent benchmarks.
Do you publish architecture details?
Not currently. Some of our work involves experimental architectures that are still under active development and are not yet publicly documented.
Are MatrixLM models production-ready?
Model training, evaluation, post-training, inference, and serving infrastructure are under active development. Availability and capabilities will expand as models and infrastructure mature.
Products & access
How can I use MatrixLM?
MatrixLM provides conversational access through the current product. Additional model and developer access will be introduced as the platform develops.
Is MatrixLM free to use?
Limited access is available without a paid subscription. Users who need additional usage can add prepaid funds to their MatrixLM account where pay-as-you-go access is available.
Do I need a subscription?
No. Pay-as-you-go usage does not require a monthly subscription. Subscription options may be introduced separately for users who want higher included capacity or additional features.
What is Pro?
Pro is a planned subscription option for users who need greater included usage and additional product capabilities. Final features and availability may change before launch.
Can companies use MatrixLM?
Business and developer access are part of the platform direction. Availability will expand as the models and serving infrastructure mature.
Billing & developer access
How does pay-as-you-go billing work?
MatrixLM uses a prepaid account balance for eligible usage. You add funds to your account and usage is deducted from your available balance as you use the service.
What is the minimum top-up?
The current minimum predefined top-up is €5. Available top-up amounts and payment options may change as billing develops.
Does my balance expire?
Any expiration rules or other conditions applying to prepaid balances will be shown in the applicable billing terms before they apply.
Will there be a developer API?
Developer API access is planned so applications and other software can use MatrixLM models programmatically. Availability, model selection, limits, and documentation will be published as the API becomes available.
How will API usage be priced?
Developer access is planned around usage-based pricing. Exact model-specific rates will be published when production API access and serving capacity are available.
Will the API support streaming?
Streaming inference is planned for developer access. Final API capabilities will be documented when developer access becomes available.
Privacy & data
Are my conversations used to train MatrixLM models?
Private user conversations are not used to train MatrixLM models unless you explicitly choose to contribute that content for that purpose.
Does MatrixLM sell personal information?
No. MatrixLM does not sell personal information or use personal information to create third-party advertising profiles.
Where is MatrixLM operated?
MatrixLM is operated from Belgium with a preference for European infrastructure and data minimization. More detailed information about processing and international transfers is provided in the Privacy Policy.
Can I request access to or deletion of my data?
You may have rights to access, correct, delete, restrict, object to the processing of, or request portability of certain personal information, subject to applicable law. Details and contact information are available in the Privacy Policy.
Where can I read the full privacy information?
The MatrixLM Privacy Policy explains what information is processed, why it is processed, retention, service providers, international transfers, and your data-protection rights.
MatrixLM
Want to learn more?
Read more about MatrixLM, explore current pricing, or review how we handle personal information.