Thox-Pro-8B — Privacy-First Professional Configuration with Vision (Ministral 3 8B, THOX-Configured)

Your AI. Your Data. Your Rules.™

Thox-Pro-8B is a THOX configuration and distribution of Mistral AI’s Ministral 3 8B: a multimodal (text and image) model that runs locally on a single device. The weight layer in this model’s Ollama manifest is byte-identical to the upstream provider’s Ollama weight layer (ministral-3:8b). THOX’s work in this release is the MeshStack clustering layer these configurations were made for, configuration (runtime parameters and, where present, a system prompt), packaging, and distribution. THOX did not train these weights.

Correction notice. This page previously presented HIPAA, SOC 2, and FERPA compliance as product properties; a locally run model can be one component of a compliant deployment, but THOX claims no certification for this model, and those checklists have been removed. It also described the build as “full precision”; the served build is Q4_K_M, per the manifest config. No benchmark, throughput, latency, concurrency, uptime, cost-saving, or compliance figure is claimed for this release; figures of that kind that appeared on earlier versions of this page were inherited catalog copy and have been removed pending THOX-run evaluation on THOX deployments. Earlier versions also stated a blanket MIT license for the model; that was incorrect for the weights (see License).

Run it

Ollama (published model)

ollama pull Thox-ai/thox-pro-8b
ollama run Thox-ai/thox-pro-8b

Summary

A THOX configuration and distribution of Ministral 3 8B for professional work on a single device: text and image input for documents, charts, and diagrams, within the base model’s abilities, with all inference local. Configuration-derived: THOX did not train these weights. This distribution’s weight layer is the same artifact as thox-cluster-swift’s (identical SHA-256 digest); the two differ in configuration and presentation. Pro 8B is packaged for single-device use, Cluster Swift for multi-device team serving.

The published Ollama model Thox-ai/thox-pro-8b is 6.0 GB. Multilingual, per the base model record (32+ languages claimed by the original catalog for the base family).

Intended use

Local inference through Ollama for professional work, on hardware that fits an 8B Q4-class model:

• Professional document and image analysis on local hardware
• Multilingual assistance for distributed teams
• Privacy-sensitive workflows where data must stay on the device, subject to your own compliance obligations

All inference local: no data leaves the device unless you send it. Downloading this model gives you local inference in Ollama. Distributed orchestration across devices is provided by MeshStack; a model download alone does not create a cluster or pool memory across devices.

Training data and procedure

No training by THOX. Base: mistralai/Ministral-3-8B-Instruct-2512 by Mistral AI, via Ollama tag ministral-3:8b. The original THOX release materials for this model claim no separate THOX fine-tuning run, training dataset, or training procedure, and none is claimed here. Base-model training data and procedure are those of the base model’s provider; see the base model card for that record. Datasets added by THOX for this release: none.

Evaluation

No THOX-run evaluation has been published for this model. The original card reports qualitative performance without published benchmark scores, and THOX claims no figure for this release. This section will be updated with method and results when a THOX-run evaluation lands.

Hardware Reference (from the original card)

Reference configuration from the original card: minimum 16 GB RAM, 8 GB storage, NVIDIA Jetson Orin NX class GPU (8GB VRAM). Recommended 32 GB RAM, 16 GB SSD, Jetson Orin AGX class GPU.

Limitations

• Outputs can be wrong. Professional, medical, legal, financial, and security-relevant outputs require qualified human review.
• Local deployment only in this distribution; behavior and speed depend entirely on the operator’s hardware and configuration.
• The operator is responsible for meeting any regulatory obligations that apply to their deployment.
• Vision is optimized for documents, charts, and professional imagery, not general photography. Maximum 5 images per request.
• All processing is local by design; the model has no access to online information or external services unless the operator wires them in.

License

The model weights are distributed under the Apache License 2.0 (license layer in this model’s Ollama manifest, from the upstream ministral-3:8b distribution). The manifest also carries an MIT license layer (Copyright 2025-2026 Thox.ai LLC) covering THOX’s own packaging material only. Earlier versions of this page stated a blanket MIT license for the model; that was incorrect for the weights. THOX claims no license over the weights beyond the upstream terms.

MagStack and MeshStack technology is proprietary to Thox.ai LLC. Publication of this model does not license THOX mesh technology, patents, or trademarks beyond identifying the model.

Provenance

THOX Pro family, the single-device professional line released alongside the MeshStack Cluster family. First listed on Ollama under Thox-ai in 2025 and held private; publicly released 2026-10-31 by order of Tommy Xaypanya, CTO and Co-Founder of THOX.ai, as part of the MeshStack model family release. Upstream: mistralai/Ministral-3-8B-Instruct-2512, Ollama tag ministral-3:8b; the weight layer is byte-identical to the upstream provider’s Ollama weight layer (SHA-256 verified in the 2026-10-31 release review) and the same artifact as thox-cluster-swift’s weight layer. Model card: https://huggingface.co/Thox-ai/thox-pro-8b. Correction history: HIPAA/SOC 2/FERPA compliance checklists removed (no certification claimed), the “full precision” description corrected to Q4_K_M, and the blanket MIT license statement corrected, as recorded in the correction notice above. Enterprise: enterprise@thox.ai.