The Meta-Agent Challenge shows frontier AI models rarely match human-engineered agent baselines when tasked with autonomous development, with proprietary models succeeding most often and some exhibiting cheating under pressure.
hub Canonical reference
Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents
Canonical reference. 85% of citing Pith papers cite this work as background.
42
Pith papers citing it
Background
85% of classified citations
abstract
Today's AI systems have human-designed, fixed architectures and cannot autonomously and continuously improve themselves. The advance of AI could itself be automated. If done safely, that would accelerate AI development and allow us to reap its benefits much sooner. Meta-learning can automate the discovery of novel algorithms, but is limited by first-order improvements and the human design of a suitable search space. The G\"odel machine proposed a theoretical alternative: a self-improving AI that repeatedly modifies itself in a provably beneficial manner. Unfortunately, proving that most changes are net beneficial is impossible in practice. We introduce the Darwin G\"odel Machine (DGM), a self-improving system that iteratively modifies its own code (thereby also improving its ability to modify its own codebase) and empirically validates each change using coding benchmarks. Inspired by Darwinian evolution and open-endedness research, the DGM maintains an archive of generated coding agents. It grows the archive by sampling an agent from it and using a foundation model to create a new, interesting, version of the sampled agent. This open-ended exploration forms a growing tree of diverse, high-quality agents and allows the parallel exploration of many different paths through the search space. Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%. Furthermore, the DGM significantly outperforms baselines without self-improvement or open-ended exploration. All experiments were done with safety precautions (e.g., sandboxing, human oversight). The DGM is a significant step toward self-improving AI, capable of gathering its own stepping stones along paths that unfold into endless innovation.
hub tools
citation-role summary
background 11
dataset 1
other 1
citation-polarity summary
representative citing papers
SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
Framework estimates context-dependent marginal utility of candidate skills via reward gaps in matched base vs. skill-augmented rollouts to filter skills and co-train policy as generator.
PACE is a training-free anytime-valid commit gate using testing-by-betting e-processes that controls per-candidate false-commit probability for self-evolving agents and reduces spurious edits compared to greedy acceptance.
MobEvolve is an agentic self-evolving heuristic framework that generates interpretable human mobility trajectories and outperforms deep generative and LLM-based methods on Singapore and Montreal benchmarks.
CyberEvolver introduces a four-layer self-evolving agent architecture with trace-to-diagnosis and population beam search that raises seed agent success rates by 13.6% on CTF, exploitation, and penetration tasks across four LLMs.
MOSS performs source-level self-rewriting in agent systems using failure-anchored pipelines and container-based verification, raising OpenClaw mean score from 0.25 to 0.61 in one cycle.
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.
Agentic-imodels evolves scikit-learn regressors via an autoresearch loop to jointly boost predictive performance and LLM-simulatability, improving downstream agentic data science tasks by up to 73% on the BLADE benchmark.
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
Comet-H orchestrates LLMs via deficit-scoring prompt selection and half-life task tracking to co-evolve research software components, demonstrated by a static analysis tool reaching F1=0.768 versus a 0.364 baseline.
AI coding agents evolve simple ground-state protocols into improved versions for VQE, DMRG, and AFQMC on spin models and molecules by using executable energy scores under fixed compute budgets.
SWE-EVO shows GPT-5.4 with OpenHands reaching only 25% success on complex multi-file evolution tasks versus 72.8% on SWE-Bench Verified, and introduces Fix Rate as a partial-progress metric.
Proposes a structured concept-centric memory system for embodied agents that connects object, scene, transition, and skill memories to support coarse-to-fine retrieval and improve task performance over baselines.
Symbolon learns diverse code transformations via search on small programs, distills them into agent skills, and applies them to improve KLEE symbolic execution, yielding 3.69x coverage gains and 21 new Linux kernel bugs.
Evolution Fine-Tuning trains LLMs on 156K trajectories spanning 371 tasks to achieve 10.22% average improvement on 22 held-out optimization tasks and match SOTA on select circle-packing problems when combined with test-time RL.
Heuresis evaluates six search strategies for autonomous ML research agents and finds that novel ideas are rare, none rated original, and only one reaches top-10 quality while strategies steer axes but do not expand the quality-novelty frontier.
AFTER benchmark shows single refinement improves LLM agent performance by 3.7-6.7 points and multi-model procedural skills reach 73.1% cross-model accuracy on 382 tasks.
Arbor combines a coordinator, executors, and a hypothesis tree to enable cumulative autonomous research, outperforming Codex and Claude Code by over 2.5x on six real tasks and reaching 86.36% Any Medal on MLE-Bench Lite.
EvoTrainer co-evolves LLM policies and training harnesses via empirical feedback to match or exceed human-engineered RL on math reasoning, code generation, and long-horizon software engineering.
MemPro evolves the entire MCR pipeline as runnable programs via failure-guided refinement on a version tree and outperforms static baselines on LongMemEval, LoCoMo, HotpotQA, and NarrativeQA.
A co-evolutionary VLM-VGM loop on 500 unlabeled images raises planner success by 30 points and simulator success by 48 percent while beating fully supervised baselines.
Shepherd provides a reversible execution trace substrate for LLM agents that enables meta-agents to inspect and transform runs, yielding reported gains on coding and terminal benchmarks via supervision, counterfactual repair, and RL credit assignment.
Generate-Select-Refine is an open-ended Bayesian optimization method that generates tasks and concentrates evaluations on the best one with only logarithmic regret overhead relative to standard single-task optimization.
citing papers explorer
-
The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development?
The Meta-Agent Challenge shows frontier AI models rarely match human-engineered agent baselines when tasked with autonomous development, with proprietary models succeeding most often and some exhibiting cheating under pressure.
-
SimWorld Studio: Automatic Environment Generation with Evolving Coding Agent for Embodied Agent Learning
SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
-
Co-Evolving Skill Generation and Policy Optimization
Framework estimates context-dependent marginal utility of candidate skills via reward gaps in matched base vs. skill-augmented rollouts to filter skills and co-train policy as generator.
-
PACE: Anytime-Valid Acceptance Tests for Self-Evolving Agents
PACE is a training-free anytime-valid commit gate using testing-by-betting e-processes that controls per-candidate false-commit probability for self-evolving agents and reduces spurious edits compared to greedy acceptance.
-
MobEvolve: An Agentic Self-Evolving Heuristic System for Interpretable Human Mobility Generation
MobEvolve is an agentic self-evolving heuristic framework that generates interpretable human mobility trajectories and outperforms deep generative and LLM-based methods on Singapore and Montreal benchmarks.
-
CyberEvolver: Structured Self-Evolution for Cybersecurity Agents On the Fly
CyberEvolver introduces a four-layer self-evolving agent architecture with trace-to-diagnosis and population beam search that raises seed agent success rates by 13.6% on CTF, exploitation, and penetration tasks across four LLMs.
-
MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems
MOSS performs source-level self-rewriting in agent systems using failure-anchored pipelines and container-based verification, raising OpenClaw mean score from 0.25 to 0.61 in one cycle.
-
Harnessing Agentic Evolution
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.
-
Agentic-imodels: Evolving agentic interpretability tools via autoresearch
Agentic-imodels evolves scikit-learn regressors via an autoresearch loop to jointly boost predictive performance and LLM-simulatability, improving downstream agentic data science tasks by up to 73% on the BLADE benchmark.
-
BIM Information Extraction Through LLM-based Adaptive Exploration
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
-
Theory Under Construction: Orchestrating Language Models for Research Software Where the Specification Evolves
Comet-H orchestrates LLMs via deficit-scoring prompt selection and half-life task tracking to co-evolve research software components, demonstrated by a static analysis tool reaching F1=0.768 versus a 0.364 baseline.
-
Optimizing ground state preparation protocols with autoresearch
AI coding agents evolve simple ground-state protocols into improved versions for VQE, DMRG, and AFQMC on spin models and molecules by using executable energy scores under fixed compute budgets.
-
SWE-EVO: Benchmarking Coding Agents in Long-Horizon Software Evolution Scenarios
SWE-EVO shows GPT-5.4 with OpenHands reaching only 25% success on complex multi-file evolution tasks versus 72.8% on SWE-Bench Verified, and introduces Fix Rate as a partial-progress metric.
-
Analytic Concept-Centric Memory for Agentic Embodied Manipulation
Proposes a structured concept-centric memory system for embodied agents that connects object, scene, transition, and skill memories to support coarse-to-fine retrieval and improve task performance over baselines.
-
Symbolon: Symbolic Execution by Learning Code Transformation
Symbolon learns diverse code transformations via search on small programs, distills them into agent skills, and applies them to improve KLEE symbolic execution, yielding 3.69x coverage gains and 21 new Linux kernel bugs.
-
Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks
Evolution Fine-Tuning trains LLMs on 156K trajectories spanning 371 tasks to achieve 10.22% average improvement on 22 held-out optimization tasks and match SOTA on select circle-packing problems when combined with test-time RL.
-
Heuresis: Search Strategies for Autonomous AI Research Agents Across Quality, Diversity and Novelty
Heuresis evaluates six search strategies for autonomous ML research agents and finds that novel ideas are rare, none rated original, and only one reaches top-10 quality while strategies steer axes but do not expand the quality-novelty frontier.
-
Managing Procedural Memory in LLM Agents: Control, Adaptation, and Evaluation
AFTER benchmark shows single refinement improves LLM agent performance by 3.7-6.7 points and multi-model procedural skills reach 73.1% cross-model accuracy on 382 tasks.
-
Toward Generalist Autonomous Research via Hypothesis-Tree Refinement
Arbor combines a coordinator, executors, and a hypothesis tree to enable cumulative autonomous research, outperforming Codex and Claude Code by over 2.5x on six real tasks and reaching 86.36% Any Medal on MLE-Bench Lite.
-
EvoTrainer: Co-Evolving LLM Policies and Training Harnesses for Autonomous Agentic Reinforcement Learning
EvoTrainer co-evolves LLM policies and training harnesses via empirical feedback to match or exceed human-engineered RL on math reasoning, code generation, and long-horizon software engineering.
-
MemPro: Agentic Memory Systems as Evolvable Programs
MemPro evolves the entire MCR pipeline as runnable programs via failure-guided refinement on a version tree and outperforms static baselines on LongMemEval, LoCoMo, HotpotQA, and NarrativeQA.
-
RoboEvolve: Co-Evolving Planner-Simulator for Robotic Manipulation with Limited Data
A co-evolutionary VLM-VGM loop on 500 unlabeled images raises planner success by 30 points and simulator success by 48 percent while beating fully supervised baselines.
-
Shepherd: Enabling Programmable Meta-Agents via Reversible Agentic Execution Traces
Shepherd provides a reversible execution trace substrate for LLM agents that enables meta-agents to inspect and transform runs, yielding reported gains on coding and terminal benchmarks via supervision, counterfactual repair, and RL credit assignment.
-
Open-Ended Task Discovery via Bayesian Optimization
Generate-Select-Refine is an open-ended Bayesian optimization method that generates tasks and concentrates evaluations on the best one with only logarithmic regret overhead relative to standard single-task optimization.
-
Training LLM Agents for Spontaneous, Reward-Free Self-Evolution via World Knowledge Exploration
LLM agents trained with a task-success reward on self-generated knowledge can spontaneously explore and adapt to new environments without any rewards or instructions at inference, yielding 20% gains on web tasks and allowing a 14B model to beat Gemini-2.5-Flash.
-
AgentGA: Evolving Code Solutions in Agent-Seed Space
AgentGA optimizes agent seeds with genetic algorithms and parent-archive inheritance to improve autonomous code generation, beating a baseline on 15 of 16 Kaggle competitions.
-
AI-Driven Research for Databases
Co-evolving LLM-generated solutions with their evaluators enables discovery of novel database algorithms that outperform state-of-the-art baselines, including a query rewrite policy with up to 6.8x lower latency.
-
Self-Optimizing Multi-Agent Systems for Deep Research
Multi-agent deep research systems self-optimize prompts through self-play to match or outperform expert-crafted versions.
-
Memory in the Age of AI Agents
The paper maps agent memory research via three forms (token-level, parametric, latent), three functions (factual, experiential, working), and dynamics of formation/evolution/retrieval, plus benchmarks and future directions.
-
Differentiable Evolutionary Reinforcement Learning
DERL is a differentiable bi-level method that evolves optimal reward structures for RL policies by composing atomic primitives and using meta-gradients from validation performance.
-
ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
ShinkaEvolve improves sample efficiency in LLM-driven program evolution via parent sampling, code novelty rejection-sampling, and bandit LLM ensemble selection, achieving new SOTA circle packing with 150 samples and gains on math reasoning and competitive programming tasks.
-
Agent libOS: A Runtime Substrate for Capability-Controlled Self-Evolving LLM Agents
Agent libOS is a runtime substrate for capability-controlled self-evolving LLM agents that completed 27 deterministic tasks without unauthorized side effects while maintaining a 7% false-denial rate.
-
Evolutionary Ensemble of Agents
EvE co-evolves code solvers and guidance states via synchronous races and Elo updates, discovering a rescale-then-interpolate mechanism that enables example-count generalization in ICON.
-
Disposition Distillation at Small Scale: A Three-Arc Negative Result
Multiple standard techniques for instilling dispositions in small LMs consistently failed across five models, with initial apparent gains revealed as artifacts and cross-validation collapsing to chance.
-
Rethinking the Value of Agent-Generated Tests for LLM-Based Software Engineering Agents
Agent-generated tests mainly act as observational feedback channels and do not meaningfully improve issue resolution success in current LLM software engineering agents.
-
Agentic Safety is an Epistemic Property, Not a Behavioral One
The paper reframes agentic safety as an epistemic property defined by teachability—the capacity to preserve future corrective leverage—rather than a behavioral property of the current policy.
-
AI for Auto-Research: Roadmap & User Guide
The paper delivers a stage-by-stage roadmap for AI in research, showing reliable assistance in retrieval and tool tasks but fragility in novelty and judgment, advocating human-governed collaboration.
-
Effective Harness Engineering for Algorithm Discovery with Coding Agents
Under fixed token budget on Circle Packing, deeper per-candidate reasoning beats generating more shallow candidates, and capable models produce evaluation hacks at higher rates.
-
The Agent Use of Agent Beings: Agent Cybernetics Is the Missing Science of Foundation Agents
Agent Cybernetics reframes foundation agent design by adapting classical cybernetics laws into three engineering desiderata for reliable, long-running, self-improving agents.
-
Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systems
This survey frames foundation agents using brain-inspired modular architectures and reviews challenges in evolution, collaboration, and safety.
- Deconstructing Superintelligence: Identity, Self-Modification and Diff\'erance
- Toward Training Superintelligent Software Agents through Self-Play SWE-RL