Overview

Self-improving agents study how systems can learn from their own actions and outcomes to perform better on later tasks, from coding and web use to robotics and long-horizon games. Common approaches turn successes and failures into reusable memories or executable skills, revise prompts and agent harnesses, or update model parameters through fine-tuning and reinforcement learning. Work increasingly uses repeated trials, held-out tasks, execution-based checks, and external evaluators to test whether improvements transfer rather than merely exploit a score. The central tension is that feedback and self-authored tests can be unreliable, so genuine progress must be balanced against regression, reward hacking, safety risks, and the cost of verification.

Latest papers

198 papers31 papers in the last four weeks, down 47% on the four weeks before. 0.2% of all new papers.

Oct 26, 2026

VISTA: Active Exploration for Self-Improvement in Embodied Agents

Agentic embodied intelligence plays a foundational role in open-ended, dynamic physical environments. Rather than treating policy learning as an end-to-end mapping from instructions to actions, recent studies reveal that embodied systems depend not only on pre-trained foundation models, but also on harness. Prior approaches largely treat rollouts as static logs for post-hoc reflection, producing harness updates without autonomously testing and refining them through subsequent embodied interaction. To address this challenge, we propose VISTA, a meta-agent framework that transforms post-hoc rollout reflection into active self-exploration for embodied harness improvement. Specifically, VISTA maintains a harness graph that records the lineage of proposed harnesses, which enables the meta-agent to retrieve prior hypotheses, build on effective mechanisms, and avoid repeatedly exploring failed directions. Each iteration begins from a retained harness version and its execution evidence. Rather than directly committing an update from these historical trajectories, the meta-agent actively explores in the environment. It autonomously generates new rollouts, probes execution failures and recovery behaviors, and iteratively modifies the harness based on the resulting evidence. The refined harness is then evaluated afresh and added to the harness graph. We evaluate VISTA on the LIBERO-Pro, RoboCasa, and RoboTwin benchmarks. Experimental results demonstrate that VISTA progressively improves task success and is more rollout-efficient than static reflection, achieving higher success rates with fewer simulator rollouts.
Oct 24, 2026

From Intelligence to Autonomy: Foundations and Construction of Autonomous Agents for Future Intelligent Systems

Artificial intelligence is moving from one-off perception and generation tasks toward long-term planning, tool use, and interaction with real environments. Greater task competence, however, does not automatically yield autonomy: an external actor may still have to set goals, sustain operating conditions, detect failures, and approve updates. We define an autonomous agent as a system that, within explicit identity and authority boundaries, maintains both task goals and viability conditions, acts continuously using models of the world and itself, and adjusts its control mechanisms and long-term capabilities through feedback. We trace the conceptual path from homeostasis and feedback control to autopoiesis and biological autonomy. We then distinguish six lines of operationalization: computational autopoiesis, behavior-based robotics, agent formalization, adaptive software, embodied adaptation, and robotic self-modeling. On this basis, we propose a one-boundary, four-loop framework in which Operational, Viability, Reflexive, and Learning loops organize task execution, self-maintenance, metacontrol, and governed learning. Five paired simulation experiments test permission constraints, road-closure response, health maintenance, capability re-estimation, and cross-task learning. Their results provide preliminary evidence for the feasibility of these mechanisms, while broader environments and physical systems remain to be tested.
Oct 22, 2026

Don't Forget What You Just Learned: Self-Directed Repair Turns Agent Mistakes Into Lasting Lessons

Intelligent agents should learn from their mistakes so that experience leads to continuous improvement. Yet today's agents recover from a mistake only temporarily: in-context reflections are easily forgotten across sessions and compete for limited context, while full-parameter fine-tuning is expensive and can interfere with existing capabilities. We introduce Self-Direction, a mechanism for agents to learn from mistakes that combines the benefits of temporary in-context retries and policy updates. When an agent fails a task, it reflects on its own trajectory in order to extract a lesson and immediately distills it into a lightweight low-rank adapter. By keeping the original weights frozen, Self-Direction prevents adapter training from overwriting base-model parameters while training additional parameters equivalent to ~0.02% of the base model's parameter count for each adapter. The agent then saves these adapters in a persistent library of lessons, automatically routing them to solve related problems later. Experiments with agents based on open-weight models from the Qwen family across open-ended research (GAIA), tool-use workflows (Claw-Eval), and terminal tasks (Terminal-Bench 2.1) show that our method improves overall success rates by 8.9 to 17.0 percentage points and performs competitively with frontier systems. We find that Self-Direction reduces repeated mistakes across future encounters of a task, nearly doubling the reliability of an agent's corrections. Learned lessons also transfer: a handful of adapters generalize to dozens of new tasks without additional training. Ultimately, this self-directed repair transforms everyday mistakes into lasting lessons.
Oct 20, 2026

Self-Improving Agents with World Model Predictive Paths

Large language model (LLM) agents are increasingly expected to interact with environments and improve their decisions during deployment. However, how anticipating future outcomes and drawing on accumulated interaction experience can jointly support decision-making remains underexplored. We propose a test-time framework, inspired by Model Predictive Path Integral (MPPI) control, that combines multi-branch predictive analysis with experience-based adaptation. At each step, the agent proposes multiple candidate actions and imagines their outcomes with a world model, so that alternatives can be compared without being executed in the environment. Since language environments provide neither explicit dynamics nor costs, the predicted outcomes are evaluated for task progress and risk, and MPPI-style weights over the resulting costs guide the final decision. A dual-branch experience memory enables self-improvement: the planner branch refines candidate generation and action selection, while the world model branch improves next-observation prediction. Experiments on ALFWorld and ScienceWorld across three backbones demonstrate the strong overall performance of our method and the effectiveness of experience-guided predictive control. Our code will be made publicly available.
Oct 20, 2026

Recoverability and Verifier Strength: Two Measured Preconditions for Unattended Agent Self-Improvement

Agent platforms are starting to improve their own agents: they read a deployed agent's execution traces, derive a change, have an automated verifier approve it, and release it with no human author. How much of this loop can safely run without a human? We propose two quantities a platform can measure. Verifier strength σσ is how well the available verifier -- often an LLM judge -- agrees with a trusted reference verifier; it limits how much a step can improve the agent. Recoverability is how cheaply a bad change can be undone: whether undoing it needs a new verifier approval, and how much of the old behavior a rollback restores. A step should run unattended only if both are adequate. On our platform this rule allows prompt, tool and skill edits to run unattended and keeps weight updates under human sign-off. Our main contribution is these two definitions and how to measure them. We built a platform that applies the rule and tested it in two ways: on the platform's own approval gate, and in controlled experiments on the τ2τ^2-bench agent benchmark. On verifier strength, the platform's LLM judge sometimes gave the same score no matter what it was shown (9 of 37 settings) and reported no error, so a verifier can look healthy while approving anything; its strength must be measured, and two simple tests catch the failure. On recoverability, restoring a saved copy brought the agent back to its original behavior, but a regenerated undo did not, so only changes whose old version is kept can be reversed without a human; and while a change was deployed, it altered the outcome of 17 percent of tasks beyond normal run-to-run variation, which no undo repairs. Together these support the rule: a change can run unattended only when its old version is kept and its verifier's strength has been measured.
Oct 19, 2026

Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy

In this paper, we systematically investigate the mechanisms, scaling behavior, and generalization of physical Recursive self-improvement. We introduce PhysRSI-eval to quantify physical RSI through a five-axis metric: First-Attempt Performance, Efficiency, Success, Budget and Improvement from failure. Different VLMs are evaluated on 18 manipulation tasks under a fixed interaction budget. The analysis of experimental results motivates the Physical Barrel Hypothesis: different capabilities may constrain different parts of improvement. The five-axis metrics also distinguish systems with close public benchmark averages and show that high scores need not come with early or efficient gains. We study observational scaling from two perspectives: predicting the relative performance of VLM agents from their public capabilities, and forecasting later progress from early physical feedback. Together, these analyses ask how quickly and how far an agent can improve. Moreover, we conduct workspace interventions and then analyze what improvement leaves behind. Removing history reduces late-round success, frozen workspaces remain useful on unseen layouts, and three cross-task cases achieve success earlier. The results suggest that scene-specific adaptation and reusable experience can coexist.
Oct 19, 2026

VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement

Recursive self-improvement (RSI) requires agents to diagnose system bottlenecks and coordinate improvements across training and inference. We refer to the agents' capability to jointly optimize training data, the trainer, and the inference harness as cross-component co-design. We introduce VibeTraining-Bench, a benchmark for evaluating this capability across tasks spanning terminal operation, software engineering, multi-turn tool use, and mathematical reasoning. Within each task, research agents start from an identical initial agent system under a fixed resource budget. Guided by validation feedback, they iteratively refine data, trainers, and inference harnesses, then select and submit a final agent system for evaluation on the test set. To prevent reward hacking, we enforce three safeguards: test data isolation, separate research and grading containers, and network restrictions for research agents. Our evaluation results show that frontier research agents substantially improve target agent systems. For example, on SWE-bench Verified, GPT-6-Astra raises the target agent system's performance from 32.9% to 53.7%, comparable to the human-developed instruct model's 55.4%. However, there is still room to improve their ability to co-design and iteratively refine data, trainer, and harness to achieve cumulative performance gains. We hope that VibeTraining-Bench will serve as a unified platform for advancing RSI, deepening our understanding of cross-component optimization for agent systems. Code and data are available at https://github.com/D2I-ai/VibeTraining-Bench.
Oct 18, 2026

DAGLoop: Aligning Method Design with Experimental Feedback in Automated AI Research

Automated AI research agents aim to improve machine learning methods by iteratively proposing, implementing, and evaluating candidate methods. Existing work uses the outcome of each evaluation to determine how the next candidate should be revised for performance improvement. However, a candidate often combines multiple design decisions, and its overall outcome alone does not reveal which decisions produced their expected effects, whether their required conditions were satisfied, or which decisions should be retained or revised. Inspired by ablation studies in the machine learning community, we propose DAGLoop to test the expected effect of a design decision by comparing implementations with and without that decision while preserving the other decisions it requires. Specifically, DAGLoop translates each design decision's rationale into a testable expectation and pairs it with a targeted experiment, forming an experimental specification for assessing the expected effect under its required conditions. Using these specifications, we formulate ablation planning as the construction of a directed acyclic graph (DAG) from the candidate method's design: nodes associate decisions with tests, and edges encode dependencies. The DAG specifies which decisions each experiment must retain and supports joint tests when an expected effect requires several decisions to act together. Outcomes are recorded with the tested decisions, expectations, and conditions to form structured research memory. This memory guides the agent in retaining experimentally supported decisions, revising unsupported ones, and planning further experiments when the evidence is unresolved. Experiments on dozens of research tasks demonstrate substantial performance improvements achieved by DAGLoop across diverse executable research settings. Code and evaluation scripts are available at https://github.com/deng34567/DAGLoop.
Oct 18, 2026

Epi: Towards a Persistent, Self-Improving Agent for Autonomous Research

We present Épi, an autonomous research agent that sustains investigations across model sessions and evolves its methods through accumulated experience. Épi combines persistent research history, hierarchical memory, branching and independent review to maintain research continuity, learn from experiments, and explore and revisit alternative approaches. Its experiments run across session boundaries, and a fixed verifier evaluates their outputs. Épi produced 120 record-setting results, comprising 20 numerical bound records across 12 families and 100 construction records across eight families. These include an approximately 84-fold narrower interval for the nn-queens constant and a more than 3.6-million-fold narrower rigorous interval for the quartic Feigenbaum attractor dimension. Exact dual certificates establish that the Cohn-Elkies linear-programming bound cannot be sharp in dimensions 10 and 11. Beyond numerical improvements, Épi gives a computer-assisted proof of Tuza's conjecture for split graphs with an eight-vertex clique part and at most three triangle-active neighborhood types, with arbitrary multiplicities. We extend this research process to algorithmic coding with Épi-turbo, a lightweight configuration that develops executable solvers for tasks including vehicle routing, scheduling and production planning. On 30 FrontierOR tasks, Épi-turbo achieves a 5.4% higher total benchmark score than Codex at 64% of its model cost, using the same underlying model and reasoning effort. Épi is designed to support human understanding by connecting computational results to explicit methods and mathematical arguments that researchers can examine and build upon.
Oct 17, 2026

MELD: Recursive Self-Improvement via Hierarchical Coordination for Multi-Agent Scientific Discovery

Multi-agent scientific discovery with large language models raises a coupled learning problem: research assignments determine both which experiments are pursued and the experiences from which agents adapt. We introduce MELD, a unified hierarchical test-time training framework that jointly adapts a team-level Coordinator policy and agent-level individual Scientist policies within a single discovery process. The Coordinator generates a joint portfolio of search branches and research directives over a shared derivation tree, while Scientists develop hypotheses and execution plans implemented by a shared frozen Executor. Both levels receive alternating GRPO-style parameter updates under a shared smooth best-of-budget objective, allowing team organization and individual research behavior to co-evolve in a recursive self-improvement loop. Our analysis establishes objective alignment for the hierarchy, characterizes conditions for coordination gains over independent parallel discovery and the first-order benefit of private adaptation, and derives a finite-budget regret bound for adapted models. On 22 MLGym and MLE-bench tasks, MELD is best on 20 of them and achieves the highest aggregate normalized scores on MLGym, using a 4B backbone for coordination and research. Most baselines use proprietary Scientists, and all systems share one frozen Executor.
Oct 16, 2026

Self-Improving Video-to-Action Models

Video generative models are an emerging tool for embodied agents: they can serve as scalable visual planners, generating videos that depict how to complete previously unseen tasks. However, executing these plans typically requires a video-to-action or "inverse dynamics" model trained on large, human-collected video-action datasets, severely limiting scalability. We introduce Mirror, a self-supervised approach that learns video-to-action mappings through environment interactions, without prior action data or hand-crafted rewards. Mirror learns by attempting to follow a visual plan by executing its best guess of the underlying actions in the environment. Mirror then trains on the resulting video-action pairs, incrementally self-improving and following the visual planner more closely. Across Atari, OGBench, and DeepMind Control, Mirror outperforms leading learning-from-observation methods with learned video plans, and achieves near-expert performance with oracle plans. On Minecraft BASALT, it outperforms methods with access to expert actions or hand-engineered task knowledge. In Minecraft survival, it is the first method to acquire an iron pickaxe through self-improvement alone, without rewards or expert action labels. In the offline setting, with expert action labels but no environment interaction, it sets a new state of the art in obtaining diamonds. Our results suggest that advances in video models can translate into increasingly capable agents without a corresponding need for hand-crafted rewards or large collections of expert action data.
Oct 16, 2026

From Trajectories to Differentiated Skills: Cognitive Differentiation for Self-Improving Agents

Large language model (LLM) agents tackle complex tasks through reasoning and interaction with external environments. Agent skills provide persistent, transferable capabilities that guide task planning, inform contextual decisions, support tool use, and enable executable procedures. Yet existing skill-evolution methods largely organize knowledge around skill artifacts, overlooking the functional heterogeneity within individual skills and its implications for knowledge maintenance and use. Inspired by cognitive science, we propose Cognitive Differentiation in Skill Evolution (CDSE), which evolves a unified skill through four functionally distinct yet coordinated components: Meta, Experience, Tool, and Script. CDSE extracts evidence-grounded semantic gradients from successful and failed trajectories, consolidates them by functional role, and applies component-specific updates and validation with cross-component coordination and adaptive invocation. An analysis of 500 top-ranked skill packages shows that 96.8% contain knowledge spanning at least two functional roles. Across six benchmarks, three agent models, and two execution harnesses, CDSE achieves the highest scores in all 36 settings, improving over the Base agent by 7.6 percentage points and the strongest competing method in each setting by 2.6 points on average. Ablation studies further support the importance of functional differentiation and adaptive invocation.
Oct 16, 2026

Reward Engine: Reward-Centric Self-Evolution for Open-Ended Tool Agents

Open-ended tool agents can complete the same objective through different tool paths, and their quality depends jointly on the interaction trace, environment state, and final artifacts, so such tasks rarely admit an executable ground-truth reward. The central challenge is therefore to construct a reward from existing experience that exposes the current policy's failures while remaining stationary during optimization. We introduce Reward Engine, which, like a self-motivated student, first works out what success requires and then optimizes toward it. From the task and initial trajectories, it builds a RewardPack of testcases for required correctness and rubrics for graded quality, each executed as a check that returns a judgment with evidence. The frozen RewardPack then guides the agent to revise and select its outputs, and can also serve as a reward for reinforcement learning. On WorkBuddy Bench Office and Claw-Eval, the constructed reward largely agrees with the benchmark graders, not reversing their ordering on 82.3% and 80.5% of candidate pairs. Used to revise executions, it raises three-run consistency (Pass3\mathrm{Pass}^{3}) over Baseline from 45.0% to 54.0% and from 78.4% to 83.7%.
Oct 15, 2026

R2^2Orch: Recursive Self-Improvement for Agent Orchestration via Self-Play within a Skill-Internalization Loop

In recent years, agentic orchestration has been widely applied to solve complex human tasks. However, existing orchestration learning still faces key challenges, including hand-designed curricula, advantage collapse when a task's rollouts all reach the same outcome, and unreliable evaluation signals that judge each skill by one noisy score. To address these challenges, we propose R2^2Orch, a new paradigm of Recursive Self-Play Orchestration: an inner loop of self-play that improves the orchestrator on self-chosen tasks, and an outer loop that improves how the inner loop improves. To support this paradigm, we introduce Recursive Learnability, a training objective that rewards a Proposer for choosing tasks whose rollouts keep reaching different outcomes as skills are withdrawn. Built on this objective, we further propose Skill Withdrawal, in which the outer loop withdraws each skill once a controlled evaluation within self-play shows that the orchestrator performs as well without it. Moreover, the two loops form one cycle: failures are distilled into skills, skills into parameters, and withdrawals lead the Proposer to new failures, so the orchestrator moves toward zero-skill deployment. Experimental results on twelve datasets show that R2^2Orch consistently outperforms baselines across question answering, math, code, and interactive tasks. Our code is available at https://github.com/beita6969/R2Orch.
Oct 15, 2026

MetaShare: Failure-Aware Collective Self-Improvement for Open-Ended Agents

Humans often learn fastest not by repeating success, but by remembering which failures not to repeat. Open-ended self-improving agents can share experience across lineages, as in GEA, or edit their own improvement mechanisms, as in DGM-Hyperagents, yet failed improvement attempts are rarely retained as reusable cross-branch knowledge, and search can collapse onto redundant strategies. We introduce MetaShare, a failure-aware framework that records successful and unsuccessful agent modifications together with signed performance deltas. Before each generation, MetaShare retrieves diverse high-performing strategies alongside high-confidence failures from other lineages, letting agents reuse effective mechanisms while avoiding known dead ends. A task-blind auditor periodically injects absent or underrepresented mechanism categories, while embedding-based de-ranking limits redundancy during retrieval. Multiple meta-agents propose modifications in parallel at heterogeneous temperatures behind a deterministic selector that gates applicability, scope, compilability, and size. Experiments across paper review, coding, math grading, robotics reward design, and financial audit-risk assessment show that MetaShare's median matches or exceeds DGM-Hyperagents' final performance and reaches its own best-performing agent in 1.7--2.9×\times fewer evolution iterations.
Oct 15, 2026

The Self-Improvement Problem

With the rise of large language models, agentic systems are becoming increasingly autonomous, as they write code, call tools, revise prompts, and propose modifications to their own components. This autonomy raises a fundamental question for safe self-improvement: can an automated gate like a reward model or an automated reviewer reliably decide whether a proposed update should be accepted? The difficulty is structural. The gate must be a terminating, computable procedure, while the property we ultimately care about, namely whether an update is a genuine semantic improvement, depends on potentially unbounded program behavior. We formalize this mismatch by modeling agents as programs and defining improvement in terms of a semantic score that need not be computable. We study two evaluation abstractions. The first is a certifier that approves or rejects an update. The second is a search procedure that, given access to a proposal generator, attempts to output some improving candidate if one exists. We show that, in the fully general setting, no terminating procedure can simultaneously guarantee reliability (never accepting a non-improvement) and non-conservatism (never missing an available improvement): self-referential constructions can force any such gate or searcher to fail. Randomized gates with bounded error face the same limit. We then identify a standard decidable regime used in practice (bounded evaluation and a finite proposal budget) under which exhaustive simulation yields a terminating search procedure with the desired guarantees. We also show the limit of this recovery: a chain of accepted updates can raise the bounded score in every round while the true score falls. Our results draw a boundary between unbounded, universal settings where correctness guarantees collapse and bounded settings where they can be recovered by design.
Oct 14, 2026

Evolving Together, Failing Apart: The Skill-Policy Co-evolution Trap in Self-Improving Agents

Skill-policy co-evolution, in which an agent's policy model and its skill bank improve each other during training, is a promising paradigm for self-improving agents. Current evaluations score the co-evolved policy on the final bank it trained with (the delivered bank). After deployment, however, skills may be rewritten, added or removed more frequently than the policy model is retrained, leaving the policy to operate on a drifted skill bank that neither the training objective nor existing evaluation protocols address. In this paper, we first establish, through theoretical and empirical analysis, that the fragility hidden by this blind spot is not incidental. The acceptance rule these methods share (cooperative acceptance) ties a policy's competence to the bank it co-evolved with, leaving it fragile once skills drift, regardless of the implementation; empirically, sensitivity to skill drift grows as co-evolution training proceeds. We therefore introduce the CO-evolution Drift Audit (CODA), a deployment-oriented evaluation protocol that simulates skill drift and, alongside success rate under worst-case drift, introduces two new metrics: a consistency rate under meaning-preserving edits and a resistance rate to corrupted skills. Across four co-evolution methods and five benchmarks (six settings), worst-case success is 5-13 points below the peak measured for each delivered pair, erasing 52-93% of the delivered bank's gain over the no-skill baseline. CODA thus reframes the target of co-evolution, from peak performance on the co-evolved bank to the ability to harness skills however they evolve, and provides a shared standard against which future methods can be audited.
Oct 14, 2026

AdaRSI: On-Demand Recursive Self-Improvement for Embodied Agents

Failures in long-horizon embodied tasks can arise from how a vision-language-action model (VLA) is called or from its execution capabilities. We introduce AdaRSI, an on-demand recursive self-improvement framework in which a coding agent first refines its tools and procedures, then adapts sub-steps that remain unreliable after testing alternative instructions and approaches. The agent collects demonstrations, tests their utility through temporary fine-tuning probes, and pools accepted data across tasks to train a specialist VLA that works alongside the original executor. At the end of each development round, its records are summarized into shared experience that guides subsequent diagnosis and data collection. On unseen composite tasks in RoboCasa365, AdaRSI raises success from 4.6% to 13.3%. On perturbed LIBERO-Pro 2.0 tasks, success rises from 32.9% to 49.1%. On RoboCasa365's Composite-Unseen suite, it achieves higher success than adapting every manipulation sub-step, using 34% of the GPU time and 36% of the demonstrations. On a real robot, the agent autonomously collects demonstrations and adapts the executor to recover fallen bottles, improving task completion.
Oct 14, 2026

MMLA-RSI: Recursive Self-Improvement through RTT and RTMA

An agent can improve its behavior within an unresolved problem by changing how it acts and which experience it retains. We propose MMLA-RSI, a recursive self-improvement scheme with two implementations in Memory-Mediated Learning Architecture: Reasoning-Time Training (RTT) changes the policy state ΦΦ, while Reasoning-Time Memory Adaptation (RTMA) changes authoritative memory MM. Each route also makes its adaptation method an object of improvement. A current numerical updater trains the policy proposing its successor; a current memory manager chooses the records available to the process proposing its successor. An independently admitted successor takes over the same role in the next generation. We give bounded executable rule languages, shared task/proposal interfaces, and one algorithm for either route or their composition. The RTT instance changes scalar schedules in a clipped-momentum update. The RTMA instance chooses NULL or a complete-row write with possible eviction. A fixed interpreter, trusted committer, and independent evaluator bound what can change. State versions, read closures, and conditional properties describe causal execution, accepted transitions, and finite work. Existing numerical-component studies motivate the construction. In four-candidate program selection, retaining LARC feedback updates reduces expected error by 24.65 and 36.65 percentage points for two trained initializations. In chronological continuous-integration replay, feedback binding is positive while retention increases loss. Completed SQL studies improve substantially over an unadapted substrate but do not establish the required additional online benefit. These results characterize their declared components. The two recursive routes and their joint realization are specified here; their beneficial recursive chains and discovery efficiency remain to be evaluated.
Oct 14, 2026

L3M: A Verifiable Harness and Runtime for Self-Extending and Self-Improving Agents

The lasting gains in AI have come from methods that absorb more computation, and language models already scale computation within a call. Useful work, though, happens across calls: over hours or days, through tools and external systems, among many agents and the humans they work for, and increasingly through code the agents write themselves. Whether that work compounds is enabled by the harness around it, and the harness faces a constraint the model does not: of the three participants in such a computation, the human, the model and code, only the human's attention does not scale, so the computation can grow only if the human is not in every loop. L3M is a runtime built on that premise, written in Lean 4 so that one language implements the harness, states its invariants, and is the language agents extend it in. One computation spans humans, agents, models and code, and all of its state lives in Git: a rich state trajectory -- rather than a chat conversation -- is an object the runtime can resume, rewind, fork, compare and merge. Waiting, communication, delegation and undo are primitives, not patterns the model reasons through each time. When an agent works out a procedure, it can write it once as a sealed Lean program whose type is its permission request, and the harness offers it back as a tool -- machinery for the model to use. To take the human out of the loop, L3M uses reversibility to decide where that is safe: a tool's review tier follows from how far its effects can be undone, so agents act freely where the runtime can recover. Because the machinery can change, the properties that make this safe are stated as kernel-checked theorems about the code that runs. On Terminal-Bench, an untuned L3M reaches about two thirds of its (unsafe) shell-equipped pass rate using only safe (not needing human review) tools.
Oct 13, 2026

EmbodiedEvo: Self-Evolving Coding Agent For Physical Intelligence

Coding agents can construct executable programs and revise them through feedback, offering a foundation for embodied control without target-task fine-tuning. Through a programmatic robot interface, the agent observes the scene, executes control programs, and adapts them to the resulting physical state, serving as the closed-loop policy. However, this within-episode adaptation does not automatically preserve experience across scenes. Our trajectory analysis shows that agents spend interaction and inference budgets establishing effective operations and identifying their preconditions, while fresh agents cannot reuse these discoveries. This motivates EmbodiedEvo, a self-evolving coding agent that distills physical feedback into reusable code Skills and diagnostic Lessons. Experience is acquired on one Playground scene per task, then frozen for independent agents in unseen scenes, without updating model weights. On RoboDojo-Sim, EmbodiedEvo achieves an average score of 44.48 and a success rate of 37.82%, surpassing all officially verified results. Paired tests show the largest gains on memory-dependent tasks and successful reuse of coordinated procedures, while fine contact control remains difficult. Experience from stronger models also improves weaker executors: Astra's experience raises Sol and Luna's success rates by 8.86 and 5.50 percentage points over no experience. Compared with raw interaction history, Skills and Lessons achieve comparable task performance at 53.0% lower execution cost.
Oct 13, 2026

RSIgen: Reciprocal Self-Improvement for Agentic Visual Generation

Agentic visual generation produces experience that can improve the native understanding (agent) and generation roles of a unified multimodal backbone, but task success alone does not determine which role should learn. Generator updates can also make the agent's action supervision stale. We introduce RSIgen, a task-centric reciprocal self-improvement framework linking learning allocation, supervision refresh, and recursive curriculum construction. Dynamic Responsibility Routing (DRR) uses matched interventions on evidence, inputs, and supported actions to select learning recipients and construct compatible training targets. After generator learning, RSIgen re-evaluates the relative utility of tested actions with branch inputs held fixed to refresh agent supervision; the retained roles then construct the next curriculum. Across BAGEL, Qwen-Image, and Wan-Image-7B, RSIgen improves AgentGen-Bench Full Overall9 by 3.2--3.3 points over a three-cycle SearchGen adaptation with action costs. Wan controls support allocation gains at matched role-level training exposure, refresh gains at a matched adaptation budget, and benefits from updated curriculum sources while recipients continue learning. In blind evaluation on 200 real-user requests, pooled tie-adjusted preference scores against the same baseline reach 70.0% for task completion and 64.9% for visual quality.
Oct 13, 2026

ReTerm: Recovery Training for Self-Improving Terminal Agents

Terminal agents have shown strong abilities in complex coding tasks. However, they often struggle to recover from errors during long-running execution (e.g., command errors, faulty implementations), eventually leading to task failures. To this end, we present ReTerm, a self-improving terminal agent trained by learning how to recover from its own failed terminal tasks. To support its training, we construct a customized data pipeline and produce ReTerm-10K, a high-quality dataset that provides effective recovery demonstrations for SFT cold start. Furthermore, we propose ReTerm-GRPO, an online RL algorithm that forks from the agent's own failure states to generate multiple independent self-recovery attempts. By learning from whether these attempts successfully complete the task, the agent self-improves through its own recovery experience. Built on Qwen3.5-9B, our ReTerm-9B achieves 34.8% Pass@1 on Terminal-Bench 2.0, improving over the base model by 14.6 points. Further analysis shows that our method achieves consistently higher recovery rates across diverse error categories compared with standard GRPO training. All code, data and model are released at https://github.com/twothsnd/ReTerm.
Oct 13, 2026

AReaL-RSI: Adaptive Scaling and Coordination of Recursive Self-Improvement

Scaling recursive self-improvement requires additional computation to produce useful research progress in reaching persistent objectives. Intuitively, more researcher agents can explore more hypotheses, but early evaluation scores might misrepresent their promise, shared findings could lose their applicability conditions, and useful procedures could remain confined to completed investigations. In this paper, we present AReaL-RSI, a framework for organizing a persistent research team of agents without updating backbone LLM weights. We formulate the scaling problem in terms of evaluated outcomes and whole-system resource expenditure. Concretely, a leader agent selects task-appropriate research orientations and guides ongoing investigations; conditional memory sharing distinguishes unresolved problems from supported findings and enables recipient adaptation. During recursive evolution, after each campaign, evidence-based review revises researcher orientation, knowledge, and skills for later campaigns. Experiments compare AReaL-RSI with Codex-Goal on downstream harness optimization, GPU kernel optimization, and AI research tasks using two backbones. Under the reported resource allocations, AReaL-RSI improves Terminal Bench performance by 2.2 percentage points for each backbone and increases mean KernelBench speedup by 32.2-36.8%; further scalability comparisons favor the design of AReaL-RSI over native scaling: combining experience sharing with specialized researcher roles yields 1.9-27.5% lower kernel latency than independent exploration and 1.3-33.9% lower latency than sharing alone across all seven comparable task-backbone pairs.
Oct 12, 2026

Imitation Holds a Key to Evolution: Localized Repair and Verified Skill Acquisition for Adaptive Agents

An agent can retain successful trajectories without learning which part of an experience should repair a new discrepancy. We formulate passive imitation (PI) as an architectural loop that localizes a violated contract, extracts a local goal, proposes an effect-matched skill, tests a small patch, and retains evidence after full-task verification. We instantiate this loop as a deterministic symbolic controller and analyze its finite repair-query complexity under explicit identifiability and noninterference assumptions. Across ten simulated tool worlds, PI solves all 600 cold held-out composition tasks with three allowed small trials, compared with 6.17% for a whole-trajectory control, while an equally expressive compositional repair control matches PI. On an unseen tool revision, PI achieves 70.0% success versus 24.0% without small trials and 79.0% with self-trial enumeration; an exploratory 24-task adaptation stream subsequently raises first-submission success to 100%, supporting a limited correction-and-retention mechanism rather than a unique incentive advantage or general intelligence. The CPU-only implementation, integrity checks, and reproducible results are available at https://github.com/FriendOfYishan/PI.
Oct 10, 2026

Does Failure Feedback Direct Agent Self-Modification?

Self-improving agents use failure feedback to revise their code or instructions, but gains after selection do not establish target-specific repair or generalization. We measure directedness by comparing repair of a randomly selected target failure with repair of the parent's other failures. Across 145 eligible modifications in public Darwin Gödel Machine logs, we find no detectable target-specific advantage. Controlled instruction-revision experiments with small open models show that relevant feedback can improve target repair. In a second study with a pre-specified analysis plan, target feedback raised repair by 20 percentage points for Qwen3-4B relative to matched same-category feedback; the 7-point increase for Phi-4-mini was not significant. Target repair did not reliably imply held-out improvement. Held-out gains were concentrated in settings with identifiable, repairable failure mechanisms, including planted errors and natural tool-interface failures. Benefits of edit localization and regression screening varied across settings, while eight-step revision loops and GEPA ended at similar mean held-out accuracy. Evaluations should report target repair, transfer to related failures, and held-out net change separately.
Oct 10, 2026

Reels-RSI: Program-Level Self-Improvement for Code-Rendered Explainer Videos

Explainer videos -- a worked geometry problem, why the sky is blue, how a dynasty fell -- are usually one more target for pixel video generation. We argue they are a different kind of artifact: a program (script, storyboard and per-frame animation code) that a deterministic renderer turns into frames. A program can be read, diffed, tested and audited; pixels can only be watched. Reels-RSI is an open-source reel director for any agent harness; its self-improvement happens at the program level rather than in weights or prompts. Every revision pass is logged with a code snapshot; a retrospective pairs each check finding with the diff that fixed it and proposes lessons; a lesson that a static check can catch is compiled into a rule with a must-fail and a must-pass example, so the memory is unit-tested; and a rule learned today runs over everything made before it. A bounded, benchmark-gated step lets an agent rewrite the director's playbook, kept only when blind pairwise judgments on held-out topics prefer it; nothing enters memory without a human decision. Across 111 reels -- a field study, two playbook evolutions, a 2x2 memory ablation, a weaker director, a 48-reel learning curve and a judge-reliability study -- we find that the rule learned in the field caught four latent defects in frames authored before it existed and none after; that rules are an audit for a strong director (eight unguarded reels, no violation) and a guard for a weaker one (thirty violations unguarded versus five); that lessons shorten revision loops but do not compound or move the judge; and that the first rewrite's 11--1 verdict was a position bias in our vote design (judges pick the second-shown reel 62--81% of the time), masking scripts that traded thirteen facts for craft. We document four failure modes of the preference signal, which check caught each, and the fixes. Code, rules, lessons and data are public.
Oct 9, 2026

RSI-Anything: Unlocking Recursive Self-Improvement with World Models

Recursive self-improvement (RSI) aims to transform interaction experience into persistent improvements, yet external feedback is costly and world-model predictions may become unreliable as policies evolve. We introduce RSI-Anything, a domain-general framework built on JEPA-Anything that uses predictive world models to guide action and learning-data selection. Observed feedback drives policy or workflow improvement and conditional world-model adaptation, while component and compatibility checks govern the retention of updated policy-model pairs. Across diverse computational and scientific domains, RSI-Anything demonstrates substantial gains. In code repair, with the coder fixed, success increases from 26.3% with a static verifier to 29.2% through predictive selection and 31.2% after world-model adaptation, at comparable test usage. Graphical user interface (GUI) evaluations showed higher task completion, while average actions decrease by 39.4%. Across five successive policy versions, laboratory-planning scores more than double from 21.80% to 46.01%. In reaction optimization, RSI-Anything achieves 96.15% mean best observed yield versus 82.16% for Bayesian optimization under the same 40-experiment budget. An end-to-end immune-science application further connects workflow evolution and predictive modeling with therapeutic hypothesis generation. These results demonstrate how predictive world models and retained system updates enable recursive improvement across diverse environments.
Oct 9, 2026

MC-EvoBench: Auditing Experience-Driven Self-Improvement in Minecraft Agents

A central challenge in embodied recursive self-improvement is enabling agents to convert interaction history into capabilities that persist beyond one world and transfer to later tasks. A saved memory or successful final action does not by itself establish that a reusable capability was acquired; evaluation must separate capability change from carried-over resources, unequal source exposure, and early execution failures. We construct MC-EvoBench as a controlled Minecraft benchmark with a fixed catalog of 19 registered evolution chains, including 12 primary evaluation chains and seven calibration or probe chains, together with three graph-aligned paired probe families for transfer. The benchmark represents each task as a task-state graph, defines stage-level environment verifiers, assigns matched source and target roles, resets the target world, and permits only a declared agent artifact to cross the source-target boundary. For each probe, no-prior, related, unrelated, and shuffled source conditions are evaluated under matched environment-step budgets and five seeds; we report TSR, WGP, and VMP and retain provenance records. The benchmark reveals a consistent direct-capability gap between controller and planner-backed variants, while related experience yields no stable transfer: it does not improve either tool probe, its small wood-species gain hinges on one seed, and all intervals include zero. Stage audits further show that early wood-collection and crafting bottlenecks often leave later capabilities unobserved. MC-EvoBench thus offers a controlled testbed for durable, transferable embodied capabilities. Code and data are available at https://github.com/xzw-ustc/mc-bench-open-source
Oct 9, 2026

ClawLoop: Self-Improving Agents via Harness-Native Training Loop

The rapid advancement of Large Language Models (LLMs) has driven their deployment toward general-purpose harness environments, where LLM-based agents operate in sandbox workspaces and interact directly with file systems and shells. While numerous post-training pipelines exist for tool-use agents, their counterparts for harness-based agents remain scarce and are predominantly one-shot and teacher-dependent: a stronger external model synthesizes tasks and collects trajectories, the target agent is trained once, and evaluation results are consumed by human researchers rather than fed back into the loop. Such pipelines are inherently bounded by the capability of the external model and the scope of human prior, leaving no room for self-improvement. To address these limitations, we propose ClawLoop, a framework that enables self-improving without any external teacher model and with minimal prior dependencies. ClawLoop composes three phases into a self-sustaining cycle: Task Synthesis, which generates a diverse pool of training tasks via two complementary construction routines, applies diagnosis-conditioned difficulty growth followed by both agent-free and agent-based task validation; Training, which updates the agent via SFT and/or RL; and Diagnosis and Feedback, which analyzes the agent's rollout behaviors to produce a capability analysis report that will be used as feedback for task synthesis in the next iteration. Using Qwen3.6-27B as the loop model, ClawLoop yields consistent improvements on SkillsBench, Claw-Eval-Live, and ClawGym-Bench. The iteratively refined tasks are further transferred to smaller-scale Qwen3.5-4B and 9B models with gains across all benchmarks, validating data quality independently of the loop model.