Papers
arxiv:2611.22565

MillBill: A Machining Agent with Geometric Feedback

Published on Nov 17
ยท Submitted by
DMITRII ZHEMCHUZHNIKOV
on Nov 18
Authors:
,
,
,
,
,

Abstract

Turning a CAD model into a machining program requires both process knowledge and geometric precision. Language models can generate such workflows by writing geometry and toolpath code for each part, but this makes numerical execution part of the model's task. We introduce MillBill, an agent that separates machining decisions from their geometric execution and verification. Given a boundary-representation part, stock, and material, a locally hosted, mid-sized open model proposes operations and cutting parameters for indexed three-axis machining. A geometric harness generates toolpaths, simulates material removal, and returns feedback for plan refinement. Setup planning accounts for what the available cutters can physically reach, rather than visibility alone. We evaluate the resulting programs by the fraction of the target surface area within the specified tolerance after simulated machining. With Qwen3.8-27B, MillBill outperforms frontier-model approaches that construct their own geometry and toolpath workflows on the evaluated machining tasks. Code: https://github.com/kulibinai/millbill

Community

Paper author Paper submitter

Can a language model program a CNC machine without computing a single coordinate? ๐Ÿ› ๏ธ

MillBill turns a B-rep part, its stock and its material into G-code for indexed 3-axis milling. The LLM only makes the machining decisions: operations, tools, cutting parameters. A geometric harness does the geometry: it recognises features, chooses setups by what the cutters can actually reach (not just by visibility), generates every toolpath, simulates material removal and returns per-feature deviations so the model can refine the plan.

๐Ÿ“Š On 100 GrabCAD parts, every program is scored by simulation. With the open Qwen3.8-27B running locally, MillBill leaves 97.4% of the target surface within 0.1 mm. The best frontier code-writing agent (GPT-6 Astra) reaches 94.3%, Claude Opus 5.5 83.0% and CNC-Net 42.3%.

๐Ÿ“ฆ Open release: code, plus 4,294 GrabCAD parts paired with milling stocks and a simulator that scores any planner's G-code.

๐Ÿ’ป https://github.com/kulibinai/millbill
๐Ÿค— https://huggingface.co/datasets/kulibinai/millbill

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.22565 in a model README.md to link it from this page.

Datasets citing this paper 1

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.22565 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.