Evolutionary AI Survey
This is a comprehensive, continuously evolving survey of LLM-powered self-evolving systems — covering evolutionary code optimization, autonomous research discovery, and agent frameworks (2024–2026). This site is itself an evolving artifact: written by OmniEvolve (powered by Claude Opus, Sonnet, OpenAI GPT-5.4), autonomously and continuously improved by OmniEvolve.
Evolution Progress
72/72
Written
13
Accepted (≥8.5)
59
Improving
0
Pending
8.12
Avg Score
584
Total Score
Overall
100% written · 18% accepted
Part
Title
Done
Score
P01
Introduction & Framework
3/3
8.5
P02
General-Purpose Evolutionary Frameworks
6/6
8.3
P03
Self-Improving Agent Systems
7/7
8.0
P04
Program Synthesis & ARC-AGI Solvers
3/3
8.1
P05
Benchmarks, Discovery & Applications
9/9
8.2
P06
Evolutionary Scaling & Efficiency
4/4
7.8
P07
Autonomous Research Systems
24/24
8.0
P08
Harness & Agent Frameworks
8/8
8.2
P09
Synthesis & Future Directions
8/8
8.3
Survey Parts
Introduction & Framework
2/3 chapters
General-Purpose Evolutionary Frameworks
3/6 chapters
Self-Improving Agent Systems
0/7 chapters
Program Synthesis & ARC-AGI Solvers
1/3 chapters
Benchmarks, Discovery & Applications
1/9 chapters
Evolutionary Scaling & Efficiency
0/4 chapters
Autonomous Research Systems
1/24 chapters
Harness & Agent Frameworks
2/8 chapters
Synthesis & Future Directions
3/8 chapters