Pith Number

pith:FO3KV7ZQ

pith:2026:FO3KV7ZQDPOKII3BE4GNKO4AXN

not attested not anchored not stored refs resolved

Harnessing Agentic Evolution

Bang Liu, Caiyin Yang, Chenglin Wu, Jianhao Ruan, Jiayi Zhang, Jinyu Xiang, Maojia Song, Yiran Peng, Yixi Ouyang, Yongfeng Gu, Yuyu Luo, Zhiguang Han, Zhitao Wang

AEvo improves agentic evolution by having a meta-agent edit the search procedure or context using accumulated evidence as state.

arxiv:2605.13821 v1 · 2026-05-13 · cs.AI · cs.LG

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{FO3KV7ZQDPOKII3BE4GNKO4AXN}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

AEvo outperforms five evolution baselines, achieving a 26 relative improvement over the strongest baseline. Across three open-ended optimization tasks, AEvo further outperforms four evolution baselines and achieves state-of-the-art performance under the same iteration budget.

C2weakest assumption

That editing the procedure or agent context via the meta-agent will reliably steer long-horizon evolution without introducing new forms of drift or instability, and that the accumulated context provides sufficient signal for effective edits.

C3one line summary

AEvo introduces a meta-agent that edits the evolution procedure or agent context based on accumulated state, outperforming baselines by 26% relative improvement on agentic benchmarks and achieving SOTA on open-ended tasks.

References

48 extracted · 48 resolved · 18 Pith anchors

[1] GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning 2025 · arXiv:2507.19457

[2] Claude Code, 2025.https://docs.anthropic.com/en/docs/claude-code/ overview 2025

[3] Anthropic’s Original Performance Take-Home 2026

[4] An improved example for an autoconvolution inequality 2026

[5] ARC-AGI-2: A New Challenge for Frontier AI Reasoning Systems 2025 · arXiv:2505.11831

Cited by

2 papers in Pith

Latent Action Reparameterization for Efficient Agent Inference

Scalable Environments Drive Generalizable Agents

Receipt and verification

First computed	2026-05-18T02:44:15.276775Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

2bb6aaff301bdca42361270cd53b80bb57c27d9af8c34cb262684f4e6a034e51

Aliases

arxiv: 2605.13821 · arxiv_version: 2605.13821v1 · doi: 10.48550/arxiv.2605.13821 · pith_short_12: FO3KV7ZQDPOK · pith_short_16: FO3KV7ZQDPOKII3B · pith_short_8: FO3KV7ZQ

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/FO3KV7ZQDPOKII3BE4GNKO4AXN \
  | 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: 2bb6aaff301bdca42361270cd53b80bb57c27d9af8c34cb262684f4e6a034e51

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "162d308e35b3b883d7d75b5eafaf151f1767463bf9dd6178a7379df5d3f1e932",
    "cross_cats_sorted": [
      "cs.LG"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2026-05-13T17:45:16Z",
    "title_canon_sha256": "6870de1fca72884d63a7ddd2babf69eda7daea3ee841b1a863a1f4dc71eca2d6"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13821",
    "kind": "arxiv",
    "version": 1
  }
}