CalibForge: Adversarial Solver Calibration for Scaling Learnable Terminal Tasks

Hugging Face Daily Papers Papers

Summary

CalibForge is an autonomous terminal-task synthesis system that uses adversarial solver calibration to create learnable tasks for training terminal agents. It constructs 5,431 calibrated tasks and improves agent performance on Terminal-Bench2.0, SWE-bench Pro, and Doc2Repo.

Training terminal agents requires executable and verifiable tasks that are not merely solvable, but appropriately challenging for learning. Executable validation establishes feasibility, yet does not reveal how a task behaves relative to a given solver setting. In this paper, we present CalibForge, an autonomous terminal-task synthesis system that uses verified solver behavior to revise candidate tasks through adversarial solver calibration. Multi-solver calibration targets disagreement within a heterogeneous solver pool, whereas contrastive solver calibration targets a designated strong-pass/weak-fail relation; both operationalize a solver-relative learnable zone anchored in demonstrated solvability. Using CalibForge, we construct 5,431 calibrated terminal tasks. Our ablations show that both strategies yield more effective supervision than authoring and validation alone or ordinary single-solver feedback. Models trained on the full collection achieve 32.58% and 47.57% on Terminal-Bench 2.0. The largest improvements over the corresponding base model reach 24.71 percentage points on Terminal-Bench 2.0, 27.68 points on SWE-bench Pro, and 30.04 points on Doc2Repo. Together, these results support solver-relative learnability as a practical target for constructing effective and transferable agent training data.
Original Article
View Cached Full Text

Cached at: 08/07/26, 05:55 AM

Paper page - CalibForge: Adversarial Solver Calibration for Scaling Learnable Terminal Tasks

Source: https://huggingface.co/papers/2608.06352

Abstract

Trainingterminalagentsrequiresexecutableandverifiabletasksthatarenotmerelysolvable,butappropriatelychallengingforlearning.Executablevalidationestablishesfeasibility,yetdoesnotrevealhowataskbehavesrelativetoagivensolversetting.Inthispaper,wepresentCalibForge,anautonomousterminal-tasksynthesissystemthatusesverifiedsolverbehaviortorevisecandidatetasksthroughadversarialsolvercalibration.Multi-solvercalibrationtargetsdisagreementwithinaheterogeneoussolverpool,whereascontrastivesolvercalibrationtargetsadesignatedstrong-pass/weak-failrelation;bothoperationalizeasolver-relativelearnablezoneanchoredindemonstratedsolvability.UsingCalibForge,weconstruct5,431calibratedterminaltasks.Ourablationsshowthatbothstrategiesyieldmoreeffectivesupervisionthanauthoringandvalidationaloneorordinarysingle-solverfeedback.Modelstrainedonthefullcollectionachieve32.58%and47.57%onTerminal-Bench2.0.Thelargestimprovementsoverthecorrespondingbasemodelreach24.71percentagepointsonTerminal-Bench2.0,27.68pointsonSWE-benchPro,and30.04pointsonDoc2Repo.Together,theseresultssupportsolver-relativelearnabilityasapracticaltargetforconstructingeffectiveandtransferableagenttrainingdata.

View arXiv pageView PDFGitHub4Add to collection

Get this paper in your agent:

hf papers read 2608\.06352

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.06352 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.06352 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.06352 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

CLI-Universe: Towards Verifiable Task Synthesis Engine for Terminal Agents

Hugging Face Daily Papers

CLI-Universe is a synthesis engine that generates verifiable terminal-agent tasks via multi-dimensional capability taxonomy and evidence-guided research, producing a distilled dataset of 6,000 trajectories. Fine-tuning Qwen3-32B on this dataset achieves 33.4% on Terminal-Bench 2.0, setting a new state-of-the-art for open-source models at or below 32B parameters.

Terminal-World: Scaling Terminal-Agent Environments via Agent Skills

arXiv cs.CL

Terminal-World introduces a fully automated pipeline that uses agent skills to synthesize high-quality training data for terminal agents, enabling models to outperform baselines with only 1.2% of the training data. The method co-derives task instructions, environments, and teacher trajectories from skill primitives.