pith:Q4HAQCFS
IntentScore: Intent-Conditioned Action Evaluation for Computer-Use Agents
A reward model that embeds planning intent scores candidate actions for GUI agents, achieving 97.5 percent pairwise accuracy and lifting success rates by 6.9 points on unseen tasks.
arxiv:2604.05157 v2 · 2026-04-06 · cs.AI
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{Q4HAQCFSPQYDKAEL64U6WKFBGG}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
IntentScore achieves 97.5% pairwise discrimination accuracy on held-out evaluation. Deployed as a re-ranker for Agent S3 on OSWorld, an environment entirely unseen during training, IntentScore improves task success rate by 6.9 points.
That the 398K offline trajectories from three operating systems contain sufficient coverage of the action distributions and intent patterns that will appear when the model is deployed as a re-ranker for new agents on new task distributions.
IntentScore learns intent-conditioned action scores from offline GUI trajectories and raises task success by 6.9 points on an unseen agent and environment.
Receipt and verification
| First computed | 2026-05-25T02:02:15.106552Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
870e0808b27c3035008bf729eb28a13185f55d7d07c3d4388e8f9fe566d0f3be
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/Q4HAQCFSPQYDKAEL64U6WKFBGG \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 870e0808b27c3035008bf729eb28a13185f55d7d07c3d4388e8f9fe566d0f3be
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "1fd3992fcca9df7bd982c234f7abb1e51ff4569acf7c0693ab87640dff200178",
"cross_cats_sorted": [],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-04-06T20:39:30Z",
"title_canon_sha256": "e9bb77a77a4875add3f89238b977fcffd78b3ed4feea3058ca7676af8539984f"
},
"schema_version": "1.0",
"source": {
"id": "2604.05157",
"kind": "arxiv",
"version": 2
}
}