thox-pro-24b

THOX configuration of Devstral-Small-2-24B-Instruct-2512 — a THOX configuration and distribution of Devstral Small 2 24B for local software-engineering work: agentic coding workflows with function calling and tool use, as supported by the base model and your tool integration

Live on Ollama

Your AI. Your Data. Your Rules.™

Status: MODEL CARD PAGE — RUNNABLE BUILD ON OLLAMA. This Hugging Face repository is the model's public model card page and THOX packaging record; the runnable build is served from Ollama (see How to Use). THOX did not train these weights; THOX's work is the configuration, clustering layer, and documentation.

Summary

A THOX configuration and distribution of Devstral Small 2 24B for local software-engineering work: agentic coding workflows with function calling and tool use, as supported by the base model and your tool integration. Configuration-derived: THOX did not train these weights. All code stays on infrastructure you control when run locally.

Correction notice. This card previously presented the base model's published SWE-Bench Verified and Terminal Bench figures as this release's own achievements, claimed ITAR compliance, and quoted per-scenario dollar gains. The benchmark figures are Mistral's for the base model and are not THOX measurements (they are attributed as upstream figures in Evaluation below); the compliance claim was unsubstantiated; and the dollar figures have been removed. No benchmark, throughput, latency, concurrency, uptime, cost-saving, or compliance figure is claimed by THOX for this release. Earlier versions also stated the license as MIT; THOX cannot grant MIT terms over Mistral's weights, and that statement has been removed (see License).

Model ID: Thox-ai/thox-pro-24b Base model: Devstral Small 2 24B by Mistral AI Runnable build: Ollama - ollama run Thox-ai/thox-pro-24b License: Apache License 2.0 (upstream terms; see License)

Highlights

• Configuration, not training: a THOX configuration and distribution of Devstral-Small-2-24B-Instruct-2512 for local software-engineering work. THOX did not train these weights; the THOX work is the configuration, clustering layer, and documentation.
• Agentic coding posture: function calling and tool use as supported by the base model and your tool integration, with all code on infrastructure you control when run locally.
• Runnable build on Ollama: ollama run Thox-ai/thox-pro-24b. Upstream benchmark figures are attributed to Mistral in Evaluation; THOX claims no figure for this release (see the correction notice in the Summary).

Files

MeshStack Origin

This model is part of the professional model family released alongside the MeshStack Cluster family. MeshStack, THOX's device-to-device clustering technology, was built for the flagship ThoxNova devices: hardware designed to magnetically connect and interact, using custom-developed magnetic pogo pins (MagStack) to distribute models and share resources device to device. The Pro family is the single-device professional line from the same program, and the foundation the Cluster family builds on.

That physical stack is the foundation of the MeshStack app and the patent-pending over-air mesh technology THOX ships today. MeshStack is included in the THOX Kickstarter campaign (thox.ai).

What THOX Did and Did Not Do

These are THOX-configured model distributions, and the THOX work lives in the MeshStack clustering layer, configuration, documentation, and the origin program, never in the weights. The weight layers in this model's Ollama distribution are byte-identical to the upstream provider's Ollama distribution (SHA-256 verified in the 2026-10-31 release review), and the weights remain governed by the upstream license terms recorded on this card. What THOX built is the layer around them: the MagStack and MeshStack clustering design these configurations were made for, the runtime configuration, and this documentation.

Specifications

Intended use

• Full-stack, mobile, and DevOps development assistance on private codebases
• Multi-file refactoring and automated test generation with human review
• Repository-scale analysis within the served context above

Run it

ollama pull Thox-ai/thox-pro-24b
ollama run Thox-ai/thox-pro-24b

This Hugging Face repository is the model's public model card page. The runnable build is served from Ollama at https://ollama.com/Thox-ai/thox-pro-24b - the original card this page is based on lives there in full.

Platforms

Training data and procedure

This is a THOX distribution of the base model named above, packaged and documented by THOX.ai for local and MeshStack/MagStack deployment. 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

Benchmark and performance figures quoted in this card were reported by Mistral for the base model and are reproduced from the model's original Ollama model card (catalog-reported at original publication). They were not re-run for this Hugging Face publication.

Independent THOX evaluation of this model on this page: pending. This section will be updated with method and results when a THOX-run evaluation lands; until then, treat quoted figures as the original catalog's claims, not as independently verified measurements.

Catalog notes for this model: Upstream figures, reported by Mistral for the base model and reproduced from the original catalog (not THOX measurements): SWE-Bench Verified 65.8%, Terminal Bench 32.0%.

Limitations

• Best suited to code-focused tasks, not general chat

• Agentic features require proper tool integration and permission scoping

• AI-generated code requires human review and full test suites before production deployment

• The operator is responsible for meeting any regulatory obligations that apply to their deployment

• All processing is local by design; the model has no access to online information or external services unless the operator wires them in.

• Outputs can be wrong. Professional, medical, legal, financial, and security-relevant outputs require qualified human review.

Hardware Reference (from the original card)

Minimum 24 GB RAM, 20 GB storage, NVIDIA Jetson Orin AGX class GPU (16GB VRAM), 8+ CPU cores. Recommended 64 GB RAM, 50 GB NVMe SSD, Jetson Orin AGX class GPU (32GB VRAM).

Research

THOX MeshStack: Capacity-Priced Pipeline Sharding, Grant-Gated KV Transfer and Lease-Fenced Device Roles for a Private Household Mesh

THOX.ai Research technical report v1.1.0, 4 November 2026. Craig Ross and Tommy Xaypanya, THOX.ai LLC. DOI: https://doi.org/10.5281/zenodo.23969070 — Paper page: https://thox.ai/research/papers/meshstack — PDF: https://thox.ai/papers/meshstack-1.1.0.pdf

License

The Ollama manifest for this model carries no license layer. The upstream publisher, Mistral, specifies the Apache License 2.0 for Devstral Small 2 on its official model card, and those upstream terms govern the weights. THOX claims no license over the weights.

Provenance

• 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.
• Runnable build: Ollama, ollama run Thox-ai/thox-pro-24b and https://ollama.com/Thox-ai/thox-pro-24b
• This Hugging Face repository was created 2026-10-31 as the model's public model card page. It contains this card; the runnable model build is served from Ollama.
• Announcement: the MeshStack model family public release announcement (thox.ai).

Citation

@software{thoxai_thox_pro_24b_2025,
  title = {Thox.ai Pro 24B - Elite Agentic Coding AI},
  author = {Thox.ai LLC},
  year = {2025},
  url = {https://huggingface.co/Thox-ai/thox-pro-24b},
  note = {Runnable build: https://ollama.com/Thox-ai/thox-pro-24b}
}

Your AI. Your Data. Your Rules.™