AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

EvoC2F: Compiling Tool Orchestration for Efficient and Evolvable LLM Agents

Lei Wei, Qi Liu, Ruiyang Huang, Xiao Peng, Sicong Xie, Lanbo Lin, Chenhao Jiang, Yuanwu Xu, Tianyuan Yang, Jiayao Liu, Li Cai, Zhaolu Kang and 1 more

ICML 2026 · Multi-agent systems and games › LLM agents · approach: Robust and efficient agent execution

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes EvoC2F, a method that compiles LLM agent tool orchestration into constrained, evolvable structures to balance flexibility, parallelism, and reliability.

Problem

Tool-augmented LLM agents must balance planning flexibility with the reliability needed for production; existing approaches either run tools sequentially without parallelism or generate unconstrained code, which hinders optimization and verification.

Approach

Compiles tool orchestration into a constrained representation that supports parallel execution and verification, and evolves it from experience; the abstract is truncated so the precise mechanism is not stated.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add semantic plan compiler (parallel scheduling) (ablation)EvoC2F · StableToolBenchSoPR / latency: 64.9 -> 72.2 SoPR; 11.2s -> 5.5sImproved
add effect annotations (ablation)EvoC2F · StableToolBenchSoPR: 64.8 -> 72.2Improved
add verification gates for skill admission (ablation)EvoC2F · StableToolBenchSoPR / regression rate: SoPR 70.5 -> 72.2; regression rate 7.2% -> 0.8%Improved
add macro-skills (ablation)EvoC2F · StableToolBenchSoPR / tool calls: 66.1 -> 72.2 SoPR; 21.5 -> 16.8 tool callsImproved
add learned skill retrieval router (ablation)EvoC2F · StableToolBenchSoPR: 70.8 -> 72.2Improved
EvoC2F vs ReAct on latencyClaude-4-Sonnet · StableToolBenchlatency (s): 16.9 -> 5.5Improved
grow skill library over sequential tasksEvoC2F · ToolSeq-500success rate: 68.2 -> 75.0Improved
use composed skills (>=2) in a taskEvoC2F · ToolSeq-500success rate: 70.6 -> 78.4Improved
increase compiler weight lambda toward reliability (ablation)EvoC2F · StableToolBench validationSR vs latency: SR 68.5 -> 74.6; latency 5.1 -> 7.8Mixed
transfer learned skills zero-shot to ShortcutsBenchEvoC2F · ShortcutsBenchASA: 46.8 -> 49.3Improved
add safe fallback disabling harmful skills (ablation)EvoC2F transfer · ShortcutsBenchASA: 49.3 -> 52.5Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy