ThoxAgent-Edge-1.7B

Local Tool-Plan Drafting Model (Qwen3-1.7B, THOX-Configured) — thoxagent-edge-1.7b is the edge tier of the THOX Agent family: upstream Qwen3-1.7B weights packaged with a THOX agent system profile specialized for local tool-plan drafting, pl

14.52 tok/s long-form (measured, CPU) · 2.43 GB footprint at 8,192 context · Safetensors + GGUF in repo

Your AI. Your Data. Your Rules.™

This repository holds the THOX packaging for ThoxAgent thoxagent-edge-1.7b: Modelfile, release manifest, and the THOX Knowledge Pack. The runnable model is published on Ollama as thox-ai/thoxagent-edge-1.7b. The weights are upstream Qwen3 weights, byte-identical to the pinned base below; THOX configured this release with a system profile and knowledge pack. THOX did not train these weights.

Summary

thoxagent-edge-1.7b is the edge tier of the THOX Agent family: upstream Qwen3-1.7B weights packaged with a THOX agent system profile specialized for local tool-plan drafting, plus the versioned THOX Knowledge Pack. Configuration-derived: THOX did not train these weights.

Highlights

• Packaging, not training: upstream Qwen3-1.7B weights packaged with a THOX agent system profile specialized for local tool-plan drafting, plus THOX Knowledge Pack v0.1. THOX did not train these weights.
• Measured on CPU: 14.52 tok/s long-form, 2.43 GB footprint at 8,192 context (see Evaluation).
• The edge tier of the THOX Agent family; safetensors + GGUF in this repo.

Specifications

Base model	Qwen/Qwen3-1.7B
Parameters	1.7B
Formats	Ollama Modelfile
License	Apache-2.0, inherited from Qwen3; retain the Apache notice. THOX-owned elements: system profile, knowledge pack, packaging.
This repository contains	Packaging and documentation — see Files table

Files

File	Size	SHA-256	Purpose
Modelfile	1,194 B	a20b761edd10b94d2630356d2de544a837b41bc3679f3a27a9b6e6e71786295e	Ollama model definition: pinned base + THOX agent system profile
manifest.json	544 B	b8ee54b9185217a4367352e99228be59efaf90bac9da5fef9927ff5f4690695d	Release manifest: base pin, profile, knowledge pack version
thox-knowledge-v0.1.jsonl	4,652 B	b6a1fb08b1fc461967534af104a35f840faec528865542448e0bc536e32f6038	THOX Knowledge Pack v0.1 (8 verified public entries, JSONL)

Run it

Ollama (published model)

ollama run thox-ai/thoxagent-edge-1.7b

Build it yourself from this repo

hf download Thox-ai/ThoxAgent-Edge-1.7B --local-dir ./ThoxAgent-Edge-1.7B
cd ./ThoxAgent-Edge-1.7B && ollama create thoxagent-edge-1.7b -f Modelfile && ollama run thoxagent-edge-1.7b

Platforms

Platform / device	Run path
Ollama (laptop / desktop)	ollama run thox-ai/thoxagent-edge-1.7b
Hugging Face	Download the builds listed under Files from this repo (see Run it)

Intended use

Local tool-plan drafting on edge hardware: turning a user goal into a clear, step-by-step tool plan the host can authorize or reject. It drafts plans only; execution always requires host authorization. Works standalone and offline.

Training data and procedure

No training. Base: Qwen/Qwen3-1.7B via Ollama tag qwen3:1.7b, pinned digest sha256:8f68893c685c3ddff2aa3fffce2aa60a30bb2da65ca488b61fff134a4d1730e7. Customization: the THOX Agent DigitalHumans kit system profile (role-specialized) in the Modelfile, plus THOX Knowledge Pack v0.1 for local retrieval.

Evaluation

Measured 2026-10-31 on CPU, x86_64 (CPU-only execution), 8,192-token context, against the pinned base digest above.

Measure	Result
Memory footprint at 8,192 ctx	2.43 GB RAM, 0.0 GB VRAM
Long-form throughput	14.52 tok/s over 1400 generated tokens
Short-answer throughput	14.68–25.97 tok/s across the factory set
Factory questions	3 prompts run (Q1 answered (439 tokens); Q2 answered (279 tokens); Q3 answered (967 tokens))
Knowledge Pack grounding	6/8 topics answered with the verified facts from the supplied passage
Fabricated-topology trap	Declined: the model refused to confirm a live three-node MeshStack state it cannot verify
Routing question	Answered as a labeled recommendation (local-first for private data)

Limitations

A 1.7B model drafts simple plans reliably and struggles with long, branching workflows; keep plans short and validate each step in the runtime. Knowledge is limited to the supplied pack and retrieval. It has no live telemetry and cannot verify live system state. Refusal behavior is the base model's, moderated only by the system profile. External actions require runtime authorization; a recommendation from this model never grants execution permission.

License

Apache-2.0, inherited from Qwen3; retain the Apache notice. THOX-owned elements: system profile, knowledge pack, packaging.

Provenance

THOX Source-Pack Full Build Program (Team AGENT), 2026-10-31. Upstream: Qwen/Qwen3-1.7B, Ollama tag qwen3:1.7b, digest sha256:8f68893c685c3ddff2aa3fffce2aa60a30bb2da65ca488b61fff134a4d1730e7. Kit: THOX Agent DigitalHumans Model Kit (catalog/models.json).

Citation

@misc{thoxai2026thoxagentedge17b,
      title={ThoxAgent-Edge-1.7B},
      author={THOX.ai},
      year={2026},
      url={https://huggingface.co/Thox-ai/ThoxAgent-Edge-1.7B}
}