pith:F7D3V4QP
KLong: Training LLM Agent for Extremely Long-horizon Tasks
KLong shows that trajectory-splitting SFT followed by progressive RL lets a 106B agent outperform a 1T model on extremely long-horizon tasks.
arxiv:2602.17547 v3 · 2026-02-19 · cs.AI · cs.CL
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{F7D3V4QPWP7OWH3FGCYFNPWK2X}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
our proposed KLong (106B) surpasses Kimi K2 Thinking (1T) by 11.28% on PaperBench, and the performance improvement generalizes to other coding benchmarks like SWE-bench Verified and MLE-bench.
That trajectories distilled from Claude 4.5 Sonnet via the Research-Factory pipeline are high-quality and that trajectory-splitting SFT plus progressive RL stages preserve sufficient context and capability for true long-horizon generalization without hidden failure modes.
KLong trains a 106B LLM agent using trajectory-splitting SFT and progressive RL to outperform a 1T model by 11.28% on PaperBench with generalization to coding benchmarks.
Receipt and verification
| First computed | 2026-06-10T12:16:10.602892Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
2fc7baf20fb3feeb1f6530b056becad5c649aa674869b745dd8e0f955258dbef
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/F7D3V4QPWP7OWH3FGCYFNPWK2X \
| 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: 2fc7baf20fb3feeb1f6530b056becad5c649aa674869b745dd8e0f955258dbef
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "9d11ab0048ddf714e0550a92ef28b9b600db89111931cbb6ba79e1f77f1fc3d2",
"cross_cats_sorted": [
"cs.CL"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-02-19T17:01:08Z",
"title_canon_sha256": "2c3764cb9431ee8c37b47e285be808e3e1a10e56063d48b7f700fdbdbef3690f"
},
"schema_version": "1.0",
"source": {
"id": "2602.17547",
"kind": "arxiv",
"version": 3
}
}