Chenyu Zhou, Qiliang Jiang, Shuning Wu +1cs.LG cs.AI
Multi-turn agentic RL increasingly treats credit assignment as a targeting problem: given a terminal verifiable reward, per-turn methods localize credit onto the turns that mattered. We identify the structural quantity that predicts when this is the right move, the verifier information density V_d = k/C (the fraction of an agent's C-step causal chain whose per-turn correctness the verifier exposes), and show that terminal-state verifiers sit deep in a low-V_d regime where targeting is the wrong axis. In controlled shared-rollout comparisons on tau^2-bench that separate reward density from credit geometry, a continuous dense reward spread uniformly beats the sparse binary outcome reward (net-harmful on 4/5 seeds), while concentrating the same advantage on progress turns or on random turns is equally harmful: targeting is second-order. The mechanism is coverage: terminal-state verification collapses the observable signal to a single final-write turn (k=1 in 98% of rollouts) while success requires a 5-8 step chain of prerequisite tool calls. A synthetic phase boundary places the crossover at V_d* ~ 0.8, whereas measured V_d is ~0.15 on tau^2-bench and ~0.4 on BFCL V3; uniform also wins on BFCL, where a matched-concentration shuffled control is negative on 8/8 seeds. The effect reproduces across model families on ToolACE-2-8B (Delta = -0.048 over 32 pre-registered seeds; an independent 20-seed replication is itself significant), and a pre-registered matched-budget breadth sweep traces a monotone dose-response whose deficit vanishes only at full chain coverage, with a reward-to-go arm reaching full-coverage parity. Uniform redistribution is the zero-information coverage default that per-turn schemes must beat; we contribute the matched-concentration shuffled control that any targeting claim should clear.
Agentic reinforcement learning (RL) often produces irregular rollout trees with shared histories. Training root-to-leaf trajectories independently recomputes these shared prefixes. Existing systems primarily target full-attention models and lack dense, differentiable hybrid-attention execution compatible with activation recomputation. We present HARTS (Hybrid-Attention RL over Tree Structures). HARTS jointly plans microbatches, data-parallel (DP) replica assignments, and microbatch-slot schedules using non-replay compact-token work after prefix compression. For chunkwise linear attention, a linear-time algorithm coordinates chunk-boundary state recovery and replay and produces the minimum number of sequential linear-attention calls under our packed execution model. HARTS preserves the chunkwise state partitioning of trajectory-wise training: it does not repeat projections, MLP/MoE computation, or final outputs, and performs only bounded state replay for numerical alignment. Per round, HARTS batches all branches into one packed call, propagates gradients through differentiable state handoffs, supports activation recomputation, and restores per-token log-probabilities. For deterministic, no-token-drop top-$k$ MoE routing, semantic multiplicities restore MoE-objective token weights and load statistics. Existing RL objectives retain their interface. To our knowledge, HARTS is the first system to demonstrate arbitrary-rollout-tree prefix-sharing speedups on a real hybrid-attention model. On an Agentic RL workload generated from SWE-bench tasks, HARTS achieves $4.81$--$4.87\times$ forward/backward/gradient speedup with activation recomputation across multiple parallel configurations. Its numerical differences are comparable to baseline self-rerun variation, and its reward trend is similar to the baseline over the first 120 steps of $τ^3$-Bench training.
Group-relative reinforcement learning waits for sibling rollouts of the same prompt, which is costly for long and variable tool-use trajectories. Single-stream Policy Optimization (SPO) removes this dependency with a persistent prompt-level value estimate, but its recipe whitens one advantage per trajectory before optimizing a token-mean actor loss. We show that trajectory centering generally does not center the token-weighted quantity consumed by the actor, and fix the mismatch by standardizing terminal-outcome advantages under the action-token measure. We additionally organize prompt evidence by the policy event that generated it rather than learner receipt order. Across matched runs on ALFWorld at two model scales and on Math-TIR, SPO++ improves online learning efficiency over SPO. A paired ablation identifies action-token-measure normalization as the strongest tested component.
Reinforcement learning with outcome-based objectives such as GRPO enables LLM-based agents to solve complex, long-horizon tasks, yet the reusable exploration patterns embedded in interaction trajectories are largely discarded after a single policy update. Existing experience-augmented approaches retrieve historical guidance at inference time, but they apply experiences without accounting for the policy's evolving capability and create persistent dependencies on external retrieval. We propose EDGE (Experience-Distillation for Guided Exploration), a framework that treats retrieved experiences as temporary training-time scaffolds and progressively internalizes their benefits into the parametric policy. Concretely, EDGE partitions each rollout group into experience-conditioned and experience-free trajectories to estimate and admit only positive marginal gains without extra sampling, then distills the induced behavior into the base policy via a reverse-KL objective on its own empirical support. A co-evolutionary experience bank further synthesizes guidance from emerging failure modes and prunes obsolete entries as the policy evolves. Across embodied, web, and search-based QA tasks, EDGE improves over strong RL baselines by up to 12.5 points and remains effective without inference-time scaffolds or a proprietary reflector. The code is available at https://github.com/xvolcano02/EDGE.
Agentic reinforcement learning (RL) has become a critical stage in the post-training of large language models. Existing critic-free, group-relative methods estimate policy advantages from multiple rollouts, avoiding the substantial memory overhead of conventional proximal policy optimization (PPO) and achieving strong performance on long-horizon interactive tasks. Despite their success, recent studies revealed three limitations: (1) Lack explicit value generalization and effective temporal credit assignment; (2) Suffer from potential advantage collapse in long-horizon complex tasks; (3) Require a costly trade-off between sampling budget and policy performance. In this work, we propose Single-rollout Autoregressive Policy Optimization (SAPO), a low-memory and compute-efficient framework in which the policy and value functions share a single autoregressive backbone. SAPO exploits the autoregressive structure of LLMs to produce policy and value predictions at distinct causal boundaries with shared parameters, while independently optimizing the PPO objectives and auxiliary on-policy SARSA objectives. To robustly estimate the contribution of each turn, we further introduce a trajectory-level generalized advantage estimator that combines lambda-returns with batch normalization. Experiments across ALFWorld and WebShop with Qwen2.5-1.5B/7B show that SAPO trains stably and outperforms PPO and GRPO by mean +15.1 and +12.1 percentage points, respectively, while eliminating the memory cost of a separate critic model and reducing per-iteration runtime by 33.2% over PPO.
Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the harness a critical part of the agent system. Our original Agent Lightning introduced a disaggregated architecture that connects arbitrary agents to RL training through an LLM endpoint proxy, an approach later adopted by frameworks such as verl Uni-Agent, AReaL 2.0, slime, and Polar. We refer to this paradigm as harnessed agentic RL, where the deploy-time harness directly participates in model post-training. Harnessed agentic RL differs fundamentally from traditional agentic RL: the harness, rather than the training engine, owns the environment interaction loop, while the trainer observes only sequences of LLM request-response pairs. This introduces challenges in retokenization, sample merging, advantage calculation, loss normalization, and backend scheduling, which can substantially affect training stability and effectiveness. We present Agent Lightning v1.0, a lightweight framework for harnessed agentic RL implemented in approximately 3,500 lines of code. It supports arbitrary agent harnesses and serves as a practical testbed for studying these challenges. We evaluate it on instruction-following, search, and coding agents, and provide a complete reproducible pipeline for coding-agent RL. Using only 6K training examples and modest compute, RL improves Qwen3.5-9B on SWE-bench Verified from 41.8% to 56.4%, a 14.6-point absolute gain. We release the complete workflow and training scripts to facilitate reproducible research on harnessed agentic RL.
On-policy self-distillation (OPSD) gives language agents dense token-level supervision from a privileged self-teacher on the policy's own trajectories. Existing methods allocate this supervision mainly by teacher trust, but trust does not reveal whether emphasizing a token supports the current policy objective. We call this the trust-utility mismatch and introduce Influence Calibration for Self-Distillation (ICSD). For each supervised token, ICSD measures the first-order response of its importance-weighted RL surrogate contribution to a teacher-directed output perturbation. Batch-adaptive calibration converts this non-stationary signal into a bounded allocation weight while preserving the original auxiliary-loss mass within each action turn. These detached weights affect only the distillation loss and require no additional model pass. Across ALFWorld, WebShop, and Search-QA, ICSD improves all matched aggregate metrics over trust-only allocation under Group Relative Policy Optimization (GRPO) and Group-in-Group Policy Optimization (GiGPO), across two model families spanning 1.5B to 7B. At 7B, it reaches 96.1% ALFWorld success and a WebShop score of 93.1. Frozen-batch analyses show that ICSD reduces teacher-supported mass assigned to objective-opposed tokens from 60.1% to 37.8% and raises cosine compatibility with the RL gradient by 0.192. A companion repository is avail- able at https://github.com/lanqz7766/Influence-Calibration-for-On-Policy-Self-Distillation-in-Agentic-RL.
Lei Bai, Jiaqi Cao, Chiyu Chen +122cs.LG cs.CL cs.CV
Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models designed to support multimodal scientific understanding, reasoning, generation, and long-horizon tasks. The training pipeline begins with scientific multimodal pre-training over rendered scientific documents, interleaved image-text data, and diverse scientific corpora. Starting from the pretrained checkpoint, we apply a unified post-training pipeline consisting of supervised fine-tuning, scalable multi-task reinforcement learning (RL), black- and white-box agentic RL, and on-policy distillation. This pipeline is supported by practical techniques that improve rollout and training stability and efficiency, including partial rollout with off-policy correction, adaptive length regularization, online speculative decoding, robust multi-task optimization, and trace-aware experience assembly for agentic tasks. At the architecture level, Intern-S2-Preview-397B extends time series modelling from efficient long-sequence understanding to numerical forecasting, while Memory Decoder is studied as a separate memory-augmented path for rapid scientific specialization without modifying the frozen 397B backbone. Evaluations across scientific, multimodal, agentic, and general-purpose benchmarks show that Intern-S2-Preview-397B achieves competitive or leading results in multiple settings. The time series modules improve scientific signal understanding and forecasting on SciTS, while the separate Intern-MemDec-4B extension improves the Biology-Instructions average score from 56.92 to 60.32 without modifying the frozen 397B backbone.
Yifu Huo, Shunjie Xing, Chenglong Wang +8cs.LG cs.CL
Agentic reinforcement learning (RL) often suffers from delayed and sparse rewards in real-world environments. A promising solution to this challenge is credit assignment, which aims to decompose trajectory-level rewards and provide more fine-grained supervision for intermediate decisions. However, existing credit assignment approaches ignore the rich process information naturally generated during environment interaction, e.g., interaction history. We argue that such information provides valuable supervision for identifying the contribution of individual actions. To this end, we propose Environmental Feedback-based Credit Assignment (EFCA), a multi-timescale credit assignment approach for long-horizon agentic RL. EFCA complements the long-term outcome signal with two environment-grounded process signals: a short-term feedback signal that captures the immediate effect of the current action and a medium-term state-history signal that identifies ineffective patterns from recent interactions. Both signals are directly extracted from environment feedback and integrated through a return reweighting mechanism. Experiments on ALFWorld and WebShop demonstrate that EFCA consistently improves both task success and task quality over strong baselines, highlighting the effectiveness of environment-grounded multi-timescale credit assignment for long-horizon agentic RL.
Recent agentic reinforcement learning methods use hindsight to complement sparse outcome rewards. However, a completed rollout can yield many such signals, leaving their appropriate allocation across turns unclear. We introduce TRIAL, a trajectory-relative hindsight distillation framework with a unified turn-aligned scoring protocol. For each decision turn, TRIAL extracts an outcome view of that decision's realized consequence and evaluates the same response under ordinary and hindsight-conditioned contexts. The signed log-probability gap determines the direction and local strength of token-level supervision, while turn-level magnitudes are normalized jointly over the realized trajectory. The resulting allocation multipliers have an eligible-token-weighted mean of one, redistributing dense supervision across turns while fixing its average multiplier. Experiments on WebShop and ALFWorld with different backbones show that TRIAL outperforms GRPO across all eight combinations of backbone, environment, and evaluation metric, while achieving the best or tied-best performance among six methods on six of them. On WebShop with Qwen3-1.7B, TRIAL improves the success rate from 56.4% to 75.2% and the task score from 78.7% to 85.7%. Controlled ablations further show that trajectory-relative turn allocation provides substantial gains beyond those of dense hindsight distillation alone.
Zi-Han Wang, Zhengxi Lu, Zhiyuan Yao +10cs.AI cs.LG
Reinforcement learning (RL) with verifiable rewards constructs trajectory-level advantage estimates, yet it often fails to credit the few pivotal decisions that determine outcomes in long-horizon, multi-turn agentic tasks. Recent work introduces privileged self-distillation for credit assignment, providing denser supervision, but it remains unclear how such local signals should represent sequential credit. We propose AgentOPSD, a critic-free, recursive method for turn-level credit assignment in agentic reinforcement learning. AgentOPSD aggregates token-level teacher-student log-probability gaps into turn-level evidence and recursively updates a Bayesian belief state in log-odds space. This yields a principled reweighting scheme that converts sparse outcome supervision into turn-level credit signals and identifies pivotal turns through the marginal belief revision between consecutive states. The method is fully compatible with standard policy optimization and requires neither an additional critic nor extra rollouts. We evaluate AgentOPSD on ALFWorld, WebShop, and Search-QA using Qwen2.5 models at two scales (3B and 7B). AgentOPSD outperforms GRPO and strong self-distillation baselines, achieving 89.1% success on ALFWorld with Qwen2.5-7B. Ablation studies attribute the gains to turn-level aggregation and history-dependent recursive belief updates.
Text-to-image (T2I) models can produce visually compelling images, yet they remain limited on open-world tasks that require complex semantic understanding, multi-step reasoning, and the integration of external world knowledge. Existing efforts introduce agent capabilities into image generation, but they either prescribe a fixed workflow or place only a subset of the open-world image generation process under agent control. Consequently, reasoning, tool invocation, and image generation are not coordinated by a single policy. We propose ToolArtist, a fully agentic image generation model obtained by post-training a Unified Multimodal Model (UMM). ToolArtist dynamically orchestrates reasoning, external tool use, and native image generation within one unified policy. During Supervised Fine-Tuning (SFT), we equip a teacher agent with search tools alongside an image-generation tool. We then convert the collected trajectories into a UMM compatible format, where the image-generation tool is concealed while the resulting generated images are retained. During Reinforcement Learning (RL), we develop an agentic RL infrastructure for UMMs and introduce Reason-Act-Draw GRPO (RAD-GRPO), which uses complementary intent and quality rewards to jointly optimize the model. Experiments show that placing the entire open-world image-generation process under an agent policy consistently outperforms approaches with fixed pipelines or only partially agent-controlled components. We release the training data and the complete post-training infrastructure.
Large language model agents have shown promising capabilities in data-driven scientific discovery tasks, where an agent interacts with an execution environment and produces a statistical claim. Long-horizon scientific analysis remains constrained by the lack of process supervised environments over real-world scientific data. This paper introduces SciDisco, a scalable framework for training Scientific Discovery agents in process-verifiable environments. SciThèque compiles hypotheses, datasets, hidden evidence graphs, and verifiers into task environments where analytical progress can be checked during interaction. DAG-grounded trajectory synthesis uses these environments to construct verifier-filtered multi-turn demonstrations. DiscoPO then uses the environment as the source of training signal, assigning turn-level credit to actions that produce verifiable analytical evidence. Experiments show that SciDisco-14B reaches state-of-the-art on hypothesis-driven scientific data analysis benchmarks.
Yikun Li, Ting Zhang, Jiakun Liu +9cs.CR cs.AI cs.SE
Real-world vulnerabilities often span multiple functions, yet most learning-based detectors classify each function in isolation: on a sample of real CVEs, we find that 71.7% of vulnerable functions require evidence from outside the function to be classified correctly. Agentic reinforcement learning (RL) could close this gap by enabling a model to gather that evidence itself, but it lacks a reliable reward, since a reward defined on the final verdict alone can be obtained without performing any investigation. We propose VulAgentRL, an agentic RL framework for interprocedural vulnerability detection built on a Code Property Graph (CPG). The CPG serves two roles: at inference time the policy queries it for callers, callees, dataflow, and other queries, and at training time the same graph verifies the evidence the policy cites. Because every CPG node carries a persistent integer identifier, this verification is an exact comparison rather than a textual match, so the reward credits verdicts that are supported by evidence. We further initialize the policy by distilling teacher investigations, and show that this warm start is necessary, since RL cannot acquire tool-use behavior it never samples. Under a repository-level split that prevents leakage, VulAgentRL outperforms state-of-the-art baselines, including frontier models, on the strict pair-wise-correct metric while issuing fewer tool calls, and its advantage persists on an out-of-distribution corpus and under class imbalance.
Large language model (LLM) agents are increasingly trained with reinforcement learning in long-horizon, sandboxed environments. Unlike conventional RL, agentic RL couples GPU-intensive rollout engines with stateful environment containers whose actions may produce visible side effects, such as file edits, command execution, and dependency installation. A single trajectory can span many rounds of gen- eration and environment interaction, so a component failure can discard completed work or expose the model to an environment state that is inconsistent with its context. However, existing systems lack efficient and correct recovery mechanisms for this distributed execution model. This paper presents Belayer, an efficient fault-tolerant system for LLM agentic RL training. Belayer handles failures in both rollout engines and environment execution while targeting low failure-free overhead. For scoped worker-local rollout failures, Belayer equips each pre-initialized shadow worker with a selective GPU-state reuse protocol that retains independently owned weights and raw KV-arena allocations after owner and GPU health checks, reinitializes worker-local state, and rebuilds request-specific KV contents from logged token prefixes. For environment failures, Belayer introduces full checkpoint and full restore to jointly capture and restore container file-system and runtime state, and coordinates the recovered environment with the LLM context to preserve prefix consistency. An adaptive policy opportunistically overlaps full-state checkpointing with natural LLM inference bubbles when the predicted interval is long enough. Empirical results show low measured overhead during failure-free training, a worker-recovery-time reduction of up to 42 times faster compared with a full engine cold start, and 1.5 to 3.5 times faster recovery from environment failures.
In long-horizon LLM agent reinforcement learning, weak policies often repeat similar failures, producing uninformative rollout trajectories and limiting effective policy optimization. Existing skill-centric methods improve exploration by optimizing, filtering, or internalizing reusable skills. However, they remain centered on the skills themselves rather than being designed as adaptive training-time support for the evolving policy. To address this, we propose a policy-centric training paradigm that reframes skills as a dynamic training scaffold. Our framework, Pats, converts rollout groups from the latest policy into evidence cards and uses task-specific evaluation to adjust the context used in subsequent rollouts. Concrete guidance helps weak policies to complete challenging tasks. As policy improves, redundant context is revised or removed to reduce reliance on explicit guidance while preserving useful rollout variation. The policy is optimized with environmental rewards using standard RLVR, and the training scaffold is discarded at deployment. On ALFWorld and WebShop, Pats improves over strong baselines by up to 18.6%. Across seven search-augmented QA benchmarks, it remains competitive while using 32.1% fewer prompt tokens than the baseline.
Agentic reinforcement learning research is constant algorithm modification, new estimators, new pipeline stages, new rollout schemes, and in mainstream frameworks each change threads through layers of trainer, distributed backend, and rollout glue: the cost lands on the researcher at every iteration. Molt is a PyTorch-native training framework built to keep that cost small: a codebase compact and clean enough for a researcher to hold in their head, and for an AI coding assistant to read and reason about in its entirety, so the algorithm flow can be traced and changed end to end. The agent is an ordinary program, and one asynchronous loop trains multimodal and mixture-of-experts policies while never training on a token it did not generate, consistent in tokens, policy versions, and model semantics. Leanness does not cost performance: under a matched, fully asynchronous protocol, Molt is statistically comparable to a state-of-the-art Megatron-based stack. Molt is open source and provides recipes and containers at https://github.com/NVIDIA-NeMo/labs-molt.
Ryan Xu, Atlas Zhao, David Bao +1cs.LG cs.AI cs.CL cs.OS
Long-horizon rollout generation has become the dominant systems bottleneck in agentic reinforcement learning (RL). As agents interact with environments over many turns, trajectories rapidly grow to tens of thousands of tokens, making synchronous RL training increasingly constrained by rollout. We propose WAR, a workload-aware rollout system that substantially accelerates synchronous agentic RL by jointly optimizing decoding and scheduling. WAR is built on a key observation: the optimal rollout optimization strategy depends on runtime load: (1) Under low load, WAR enables model-free speculative decoding with SuffixDecoding, which reuses suffix patterns from previously completed trajectories as speculative drafts for future rollouts. Unlike model-based drafters, SuffixDecoding introduces no additional draft model and avoids GPU contention with rollout generation. (2) Under high load, where saturated batched decoding leaves limited room for speculative speedup, WAR shifts the optimization focus to cache-aware scheduling. A global scheduler places requests across rollout replicas based on cache locality, trajectory progress and server load, reducing redundant KV-cache recomputation and mitigating load imbalance. By combining decoding-level suffix reuse with system-level rollout scheduling, WAR delivers robust throughput improvements across workload regimes without changing the underlying RL algorithm. WAR improves long-context agentic rollout throughput by 1.4x under low load and up to 1.6x under high load. These results show that WAR removes a major rollout bottleneck in synchronous agentic RL and provides a practical path toward scalable long-context agent training.
Muon is competitive with AdamW in large-scale pre-training, but its operating regime in reinforcement-learning post-training remains unclear. We map this regime on ALFWorld, a sparse-reward agentic benchmark, using three group-based objectives and Qwen2.5 models from 0.5B to 3B. Under a shared KL and clipping recipe, matched optimizer comparisons and AdamW rate controls trace the usable step-size range. AdamW responds non-monotonically to rate, whereas fan-in Muon remains stable at a more aggressive effective step: at $3 \times 10^{-5}$ it improves late success over an AdamW $10^{-6}$ baseline after correction across rate-metric tests. Its normalized-AUC effect is directionally positive but less uniform; the heuristic-matched lower-rate effect is less consistent, and tuned AdamW nearly matches high-rate Muon at 3B GraphGPO. High-rate Muon applies $3.53 \times$ AdamW's hidden-matrix update RMS; a full-budget RMS-matched control removes the late-success gain. Together, these results identify a recipe-level operating regime in which fan-in Muon supports a more aggressive stable effective step under shared KL and clipping: the margin is largest when optimization headroom remains and contracts near saturation, after AdamW tuning, or under magnitude matching. The scale-matched control ties this spectral effect to Muon's scale convention rather than establishing a universal optimizer ranking. Code is available at https://github.com/x66ccff/verl-muon.
Large language models are increasingly trained as interactive agents for long-horizon tasks involving multi-turn interaction, tool use, and environment feedback. Outcome-based reinforcement learning (RL) provides a practical optimization paradigm, but its sparse trajectory-level rewards offer limited guidance on intermediate decisions, leaving a supervision gap between episode-level outcomes and token-level policy learning. We propose SEED (SElf-Evolving On-Policy Distillation), a self-evolving framework that converts completed on-policy trajectories into training-time hindsight skills and distills their behavioral effect back into the policy model. SEED first fine-tunes the policy to analyze completed trajectories and generate natural-language skills that capture reusable workflows, decisive observations, or failure-avoidance rules. During RL, the current policy both collects trajectories and serves as the analyzer that extracts hindsight skills from them. Policy updates therefore improve subsequent decision making and skill analysis together, allowing hindsight supervision to evolve with the policy. SEED then re-scores the sampled actions under ordinary and skill-augmented contexts, converting the skill-induced probability shift into a dense token-level on-policy distillation signal. This signal is jointly optimized with outcome-based RL, keeping the auxiliary supervision aligned with the current trajectory distribution. Extensive experiments on text-based and vision-based agentic tasks show that SEED consistently improves performance and sample efficiency, exhibiting robust generalization to unseen scenarios. Our code is available at https://github.com/jinyangwu/SEED.
Multi-turn agents solve complex tasks through extended sequences of tool interactions before producing a final answer, making credit assignment a fundamental challenge during post-training. Outcome rewards provide reliable supervision for short-horizon reasoning, but become sparse and high-variance as trajectories grow to tens or hundreds of tool calls. They can also be misleading: a failed rollout may contain many useful actions that move the agent closer to the goal, yet outcome-only training assigns them the same negative advantage as the eventual mistake. We propose TRACE (Turn-level Reward Assignment via Credit Estimation), a dense credit-assignment method for agentic reinforcement learning. TRACE represents rollouts as state transitions at tool-call boundaries, obtains gold-answer log-probabilities from a frozen reference model, transforms them into log-ratio state values, and derives per-action rewards as Temporal-Difference changes in those values. This requires no additional critic or process-label training, and its one-step log-ratio TD component telescopes across redundant tool calls. On long-horizon complex search, TRACE substantially improves base-model tool-use ability using pure RL, without a cold-start supervised fine-tuning stage, an agentic mid-training stage, or training on live-web data. On the closed-web BrowseComp-Plus benchmark, it raises Qwen3-4B from $7.2$ to $35.6$ and Qwen3-30B-A3B from $8.4$ to $42.6$. The learned search behavior also transfers to open-web benchmarks, and the learning curves show earlier improvement and faster convergence during RL training.
Group-based reinforcement learning (RL) has become an effective paradigm for improving large language model agents on long-horizon interactive tasks. To obtain finer-grained policy updates than trajectory-level optimization, recent work has moved toward step-level group-based RL, where intermediate steps are grouped and compared within a rollout batch. However, step-level advantage estimation is sensitive to how groups are formed: grouping by broad state keys improves coverage but may compare actions taken under different histories, while enforcing historical consistency yields fairer comparisons at the cost of fragmented groups and missing peer-comparison signal. In this paper, we propose ProGPO (Progress- and Reliability-Oriented Group Policy Optimization), a learned-critic-free method for context-consistent step-level learning. ProGPO keeps exact-prefix action comparison, and complements sparse peer comparisons with transition credit derived from rollout-based state potentials. To estimate these potentials reliably, ProGPO combines semantic expansion with inverse-variance fusion across history depths. We evaluate ProGPO on two challenging agentic tasks, ALFWorld and WebShop, with Qwen2.5-1.5B-Instruct. Results show that ProGPO improves over matched agentic RL baselines under comparable computational overhead, and additional Qwen2.5-3B-Instruct experiments further test the scalability of the proposed method.
Self-evolving frameworks usually optimize task solutions while treating the surrounding harness as fixed. We introduce Harness-Aware Self-Evolving (HASE), an agentic reinforcement-learning framework in which a single model can generate task solutions or edit selected harness components in a multi-turn action space. HASE enables a single Qwen3-8B model to match the text-classification performance of a GPT-OSS-120B model that uses Claude Code as the harness proposer. In alpha factor mining, HASE outperforms the reported GPT-OSS-120B baseline. HASE also repairs imperfect evaluation components and converges to state-of-the-art performance in circle-packing algorithm discovery. These results show that HASE improves the harness and the solution through one unified agentic process.
Yuanda Xu, Zhengze Zhou, Hejian Sang +5cs.LG cs.AI
Agentic reinforcement learning requires assigning credit to environment-facing actions such as searches, clicks, edits, navigation commands, and object interactions. Standard GRPO uses the final verifier outcome as a uniform advantage over all action tokens. This outcome signal is useful but structurally incomplete: it punishes useful exploration in failed rollouts and reinforces redundant or regressive actions in successful rollouts. We propose TRIAGE, a role-typed credit assignment framework that adds a semantic role axis to outcome credit. A structured judge classifies each segment as decisive progress, useful exploration, no-progress infrastructure, or regression, and a fixed role-conditioned rule maps these labels to bounded segment-level process rewards. This keeps verifier outcomes as the source of optimization direction while correcting the two main blind spots of outcome-only credit. We further show that role-conditioned credit is the optimal segment-level correction expressible from role labels alone -- a projection of the per-segment advantage residual onto the role variable -- so that the fixed role constants reduce advantage estimation error whenever the judge is reliable, and we connect this to lower-variance policy gradients. Across ALFWorld, Search-QA, and WebShop, TRIAGE improves success rates over GRPO for two policy models and outperforms both a scalar judge-derived process reward and an outcome-supervised shared-backbone value baseline. Ablations show that the gain comes from role typing rather than merely adding dense rewards: reliable detection of regression inside successful trajectories is the dominant contributor, while exploration credit provides a consistent secondary gain; on completed ALFWorld and WebShop rollouts, TRIAGE also reduces environment-facing turns by an additional $10.4\%$ and $14.8\%$ relative to GRPO.
Long-horizon language agents must repeatedly interact with tools, accumulate evidence, and make decisions under bounded context windows. Existing context-management methods make such rollouts feasible by truncating distant history, folding past turns into summaries, or selecting compact memory states. However, these breakthroughs introduce two coupled limitations. First, as the number of turns grows, historical observations are progressively removed or collapsed into compressed states, making it harder for the policy to reuse fine-grained evidence. Second, once the original turns are no longer source-addressable, outcome-based RL loses an explicit path for aligning policy updates with the evidence that supported a successful final answer. To this end, we propose ECHO, a selective turn-memory framework that jointly addresses history collapse and traceable learning through source-indexed reconstruction. Specifically, ECHO compresses each completed environment turn into a compact memory record, reconstructs bounded policy contexts by selecting from these records, and reuses the selected source indices to route positive outcome credit to the evidence and selection actions that support successful answers. On BrowseComp-Plus, ECHO reaches 43.4% held-out accuracy, outperforming GRPO (28.9%) and the rolling-summary baseline SUPO (36.1%), while using fewer turns and lower trajectory volume than SUPO (Figure 1). Additionally, the trained policy improves zero-shot generalization across multi-objective QA, code generation, and deep information-seeking benchmarks on both dense and MoE backbones.
Skill memories can improve agentic reinforcement learning by reusing past experience as textual guidance, but retrieved skills are not oracular: they may help in one state while misleading the same policy in another. This makes the common privileged-teacher assumption fragile, namely that a skill-conditioned prompt can be treated as a fixed teacher for the no-skill prompt. We introduce UCOB, a framework for learning to utilize and evolve agentic skills via credit-aware on-policy bidirectional self-distillation. UCOB treats skill-conditioned and no-skill prompts as two on-policy context views of the same model, compares their return-to-go within the same task and anchor state, and uses the higher-return view as the local teacher. This local credit signal internalizes useful skill-conditioned behavior, corrects misleading skill usage, and guides task/state skill memory updates, utility-aware retrieval, and reflection self-training. Experiments on agentic tasks, including ALFWorld, WebShop, and Search-QA, show that UCOB outperforms skill-free RL, skill-memory baselines, and self-distillation methods across model scales, with up to 23.5 and 18.0 point gains over SOTA baselines on ALFWorld and WebShop. Ablations and analyses further validate its core mechanisms, continual adaptation across environments, and modest training overhead. Code is available at https://github.com/TU2021/UCOB.
Self-distilled agentic reinforcement learning augments trajectory-level reward with a token-level distillation loss, using as its teacher the same policy conditioned on privileged context. The prevailing recipe gates this loss by a single scalar, the teacher-student log-probability gap. This signal is doubly limited: it is retrospective, scoring only the realised rollout and never the counterfactual ones, and it is sign-blind, never signalling when a teacher-preferred action would have harmed the trajectory. We introduce CRAFT, a three-pillar credit-assignment scheme that addresses both limitations. Pillar 1, Counterfactual Token Importance, reuses the G-1 sibling rollouts that GRPO already samples and importance-weights them by the log-probability gap to form a self-normalised estimate of the group-level counterfactual change in advantage from up-weighting teacher-preferred actions at each step; this yields a signed per-token credit at near-zero extra compute. Pillar 2 is an asymmetric controller that raises the distillation weight as it lowers the reference-KL weight along an exponential moving average of gate activity, and conversely. Pillar 3 polarises the KL penalty token by token, switching between a mode-seeking and a mode-covering update according to the sign of the credit. Each pillar has an independent switch that, when disabled, renders the loss and gradient byte-identical to the baseline in IEEE-754 arithmetic, so any measured gain is attributable to algorithmic change rather than implementation drift. We prove the estimator's consistency and a variance bound, give structural and bit-exact reproducibility guarantees, and evaluate CRAFT across three agentic environments, four model scales, and five end-to-end methods, plus two tabulated prior-work baselines. Among these is Adaptive-CRINGE, a comparator sharing Pillar 2 with CRAFT, isolating the counterfactual contribution.
A world model predicts environment dynamics based on current observations and actions, serving as a core cognitive mechanism for reasoning and planning. In this work, we investigate how world modeling based on language models can further push the boundaries of general agents. (i) We first focus on building foundation models for agentic environment simulation. We introduce Qwen-AgentWorld-35B-A3B and Qwen-AgentWorld-397B-A17B, the first language world models capable of simulating agentic environments covering 7 domains via long chain-of-thought reasoning. Leveraging more than 10M environment interaction trajectories of 7 domains in real-world environments, we develop Qwen-AgentWorld through a three-stage training pipeline: CPT injects general-purpose world modeling capabilities from the state transition dynamics and augmented professional corpora, SFT activates next-state-prediction reasoning, and RL sharpens simulation fidelity through a tailored framework with hybrid rubric-and-rule rewards. To evaluate language world models, we present AgentWorldBench, a comprehensive benchmark constructed from real-world interactions of 5 frontier models on 9 established benchmarks. Empirical results demonstrate that Qwen-AgentWorld significantly outperforms existing frontier models. (ii) Beyond foundation models, we further investigate two complementary paradigms through which world modeling enhances general agents. First, as a decoupled environment simulator, Qwen-AgentWorld supports scalable and controllable simulation of thousands of real-world environments for agentic RL, yielding gains that surpass real-environment training alone. Second, as a unified agent foundation model, world-model training acts as a highly effective warm-up that improves downstream performance across 7 agentic benchmarks. Code: https://github.com/QwenLM/Qwen-AgentWorld
Yunan Wang, Minghui Song, Zihan Zhang +6cs.LG cs.AI cs.CL
Group-based Reinforcement Learning (RL) has significantly enhanced Large Language Models (LLMs) in agentic scenarios. To achieve finer-grained policy updates, recent agentic RL frameworks have shifted from trajectory-level to step-level training. However, long-horizon agentic RL suffers from severe reward sparsity and delay, as feedback is often deferred for dozens of interaction steps. While existing step-level frameworks refine training granularity, their credit assignment remains coarse-grained and still treats agent exploration as isolated, linear trajectories. This oversimplified perspective ignores the inherent graph structure of state transitions, leading to high-variance state-value estimation and myopic, localized credit assignment. To overcome these critical bottlenecks, we propose Group-Graph Policy Optimization (G2PO), a novel group-based RL algorithm tailored for multi-turn agentic tasks. G2PO explicitly transforms linear interaction trajectories into a global state-transition graph. By aggregating identical observations across different trajectories, we introduce group-aggregation state-value estimation that reduces sampling variance and trajectory-dependent bias. Furthermore, we redefine agent actions as transitions between state nodes and propose an edge-centric advantage estimation strategy. By globally standardizing Temporal Difference (TD) errors across the entire graph, G2PO explicitly identifies and prioritizes critical transitions that drive absolute task progress. Extensive experiments on representative long-horizon benchmarks-WebShop, ALFWorld, and AppWorld-demonstrate that G2PO substantially outperforms state-of-the-art prompt-based and RL baselines, achieving remarkable success rate improvements of up to 22.2% over GRPO.
Reinforcement learning (RL) has emerged as a powerful paradigm for training Large Language Models (LLMs) as agents. However, conventional RL methods for long-horizon agentic tasks often struggle with sparse outcome rewards. Intuitively, this overlooks the rich environment dynamics information contained in rollout interaction trajectories. We argue that the interaction experience inherently serves as an implicit supervision signal, reveals the underlying transition mechanisms of the environment, and enables the agent to construct a more accurate internal model of the environment.. Therefore, in this work, we investigate how to leverage this additional signal to improve policy learning. Specifically, we propose EnvRL, a framework that incorporates environment dynamics learning into agentic RL via two auxiliary objectives: state prediction and inverse dynamics. By jointly optimizing with the primary RL objective, we encourage the agent to internalize environment dynamics from its own interaction experience. Extensive experiments on two long-horizon agentic benchmarks demonstrate that EnvRL achieves significant improvements on success-rates over RL-only baselines, e.g., when trained with GRPO, lifting Qwen-2.5-1.5B-Instruct from 72.8% to 77.4% on ALFWorld, and from 56.8% to 67.0% on WebShop.