CoworkOS:Scaling Agent Coworkers for Real-world Automation
READ THE ORIGINAL SOURCEOpen original sourcehuggingface.co · Opens in a new tabCoworkOS is presented as a general-purpose harness for long-horizon productivity work that separates task strategy from execution management. The model decides actions, delegation, and acceptance, while the runtime maintains state and feedback. A three-layer architecture supports direct execution or delegation to Subagents, synchronously or asynchronously. A shared application layer lets humans and agents edit the same documents, CAD models, and Blender scenes, and hierarchical workspaces with persistent semantic memory support resumable work. Across 12 benchmarks, the authors report competitiveness with Codex and DeepSeek Harness. In a case study using GLM-5.3 as backbone, average score on eight benchmarks for Qwen3.6-35B-A3B rose from 33.19 to 42.95 over five rounds of automated post-tra
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
- Linked in sourcehttps://github.com/RUC-NLPIR/CoworkOS
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
CoworkOS is presented as a general-purpose harness for long-horizon productivity work that separates task strategy from execution management. The model decides actions, delegation, and acceptance, while the runtime maintains state and feedback. A three-layer architecture supports direct execution or delegation to Subagents, synchronously or asynchronously. A shared application layer lets humans and agents edit the same documents, CAD models, and Blender scenes, and hierarchical workspaces with persistent semantic memory support resumable work. Across 12 benchmarks, the authors report competitiveness with Codex and DeepSeek Harness. In a case study using GLM-5.3 as backbone, average score on eight benchmarks for Qwen3.6-35B-A3B rose from 33.19 to 42.95 over five rounds of automated post-tra
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; comparisons against Codex and DeepSeek Harness rely on the authors' own benchmark setup.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- The model decides actions, delegation, and acceptance while the runtime maintains state and feedback.
- A shared application layer lets humans and agents co-edit documents, CAD models, and Blender scenes.
- Across 12 benchmarks, CoworkOS is reported as highly competitive with Codex and DeepSeek Harness.
- Over five rounds of automated post-training, an eight-benchmark average rose from 33.19 to 42.95.
Source claims not independently checked; comparisons against Codex and DeepSeek Harness rely on the authors' own benchmark setup.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
The work targets the open problem of pairing high-quality complex productivity output with monitoring over long horizons. It also reports a case where an agent harness drove automated post-training of another model.
Limits and open questions
Source claims not independently checked; comparisons against Codex and DeepSeek Harness rely on the authors' own benchmark setup.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.