Pith. sign in

REVIEW 16 cited by

AgentGym: Evolving Large Language Model-based Agents across Diverse Environments

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.04151 v1 pith:AMEYCFOF submitted 2024-06-06 cs.AI cs.CL

AgentGym: Evolving Large Language Model-based Agents across Diverse Environments

classification cs.AI cs.CL
keywords agentsenvironmentsagentgymacrossagentdiverseexplorationsuite
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Building generalist agents that can handle diverse tasks and evolve themselves across different environments is a long-term goal in the AI community. Large language models (LLMs) are considered a promising foundation to build such agents due to their generalized capabilities. Current approaches either have LLM-based agents imitate expert-provided trajectories step-by-step, requiring human supervision, which is hard to scale and limits environmental exploration; or they let agents explore and learn in isolated environments, resulting in specialist agents with limited generalization. In this paper, we take the first step towards building generally-capable LLM-based agents with self-evolution ability. We identify a trinity of ingredients: 1) diverse environments for agent exploration and learning, 2) a trajectory set to equip agents with basic capabilities and prior knowledge, and 3) an effective and scalable evolution method. We propose AgentGym, a new framework featuring a variety of environments and tasks for broad, real-time, uni-format, and concurrent agent exploration. AgentGym also includes a database with expanded instructions, a benchmark suite, and high-quality trajectories across environments. Next, we propose a novel method, AgentEvol, to investigate the potential of agent self-evolution beyond previously seen data across tasks and environments. Experimental results show that the evolved agents can achieve results comparable to SOTA models. We release the AgentGym suite, including the platform, dataset, benchmark, checkpoints, and algorithm implementations. The AgentGym suite is available on https://github.com/WooooDyy/AgentGym.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 16 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Doomed from the Start: Early Abort of LLM Agent Episodes via a Recall-Controlled Probe Cascade

    cs.AI 2026-07 conditional novelty 7.0

    A cascade of recall-calibrated gates on LLM agent hidden states aborts doomed episodes early, saving up to 47% compute at a 90% global success-recall target.

  2. A Scalable Approach to Evaluating Moral Sensitivity in LLMs

    cs.CY 2026-07 conditional novelty 6.5

    Under morally irrelevant noise, eight LLMs preserve the semantic content of identified moral features above calibrated floors, despite significant changes in feature counts.

  3. Agent Security Needs Redefinition through a Holistic Framework

    cs.CR 2026-07 conditional novelty 6.0

    Agent security should be redefined around four contextual authorization properties instead of the content of the action performed.

  4. Doomed from the Start: Early Abort of LLM Agent Episodes via a Recall-Controlled Probe Cascade

    cs.AI 2026-07 conditional novelty 6.0

    Hidden-state probes can flag doomed LLM-agent episodes from the first round, and a recall-calibrated cascade of abort gates cuts generated tokens by up to 60% while preserving a chosen success-recall target.

  5. From Trajectories to Prefixes: Reusing Teacher Trajectories via Replayed Prefixes and Online Continuation

    cs.LG 2026-07 conditional novelty 6.0

    Replaying teacher trajectory prefixes and explicitly optimizing the historical prefix tokens improves small-model agent success rates over distillation and response-only GRPO baselines in TextCraft, BabyAI, and ALFWorld.

  6. COMAP: Co-Evolving World Models and Agent Policies for LLM Agents

    cs.AI 2026-06 unverdicted novelty 6.0

    COMAP co-evolves textual world models and agent policies for LLMs through on-policy self-distillation, yielding up to 16.75% relative gains on embodied planning, web navigation, and tool-use tasks.

  7. Learning to Adapt: Self-Improving Web Agent via Cognitive-Aware Exploration

    cs.AI 2026-05 unverdicted novelty 6.0

    SCALE introduces three adversarial roles (Selector, Predictor, Judger) and a graph exploration method (SCALE-Hop) to enable MLLM-based web agents to self-discover limitations and improve, backed by the SCALE-20k datas...

  8. Test-Time Deep Thinking to Explore Implicit Rules

    cs.AI 2026-05 unverdicted novelty 6.0

    TTExplore trains a 7B thinker via task-score RL to infer implicit rules at test time, raising agent success by 14-19 points on five embodied tasks.

  9. SEAL: Synergistic Co-Evolution of Agents and Learning Environments

    cs.CL 2026-05 unverdicted novelty 6.0

    SEAL co-evolves LLM agents and environments via shared turn-level failure diagnoses, yielding +8.25 to +26.25 point gains on tool-use tasks with only 400 samples.

  10. From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails

    cs.AI 2025-10 unverdicted novelty 6.0

    Control-theoretic guardrails enable proactive correction of risky LLM agent actions in latent space, preventing catastrophes like collisions or bankruptcy while preserving task performance in simulated environments.

  11. Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents

    cs.AI 2024-08 unverdicted novelty 6.0

    Agent Q integrates MCTS-guided search, self-critique, and off-policy DPO to train LLM agents that outperform behavior cloning and reinforced fine-tuning baselines in WebShop and achieve up to 95.4% success in real-wor...

  12. NeSyFS: A Neuro-symbolic Fast-Slow Thinking Framework for LLM Agent under Partial Observability

    cs.AI 2026-07 conditional novelty 5.0

    Combining a knowledge-graph belief state, KG-augmented reflection, and TSMC-style particle planning lifts LLM agent success rates by roughly 30-95% over ReAct/Reflexion baselines on three benchmarks.

  13. AgentOmnia: Scaling Agentic Models for Full-Scenario Applications

    cs.AI 2026-07 conditional novelty 5.0

    Post-training an open Qwen3-30B agent on 52,361 verifiable tasks in 5,018 synthesized stateful environments lifts its average across four agent benchmarks from 22.9% to 41.7%.

  14. Evolutionary Intelligence for Scientific Discovery: From Evolutionary Computation to Cumulative Discovery Systems

    cs.NE 2026-07 accept novelty 5.0

    Evolutionary intelligence reframes evolutionary computation as cumulative scientific discovery by retaining search trajectories, failures, and lineages across cycles.

  15. GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

    cs.CL 2025-08 unverdicted novelty 4.0

    GLM-4.5, a 355B-parameter MoE model with hybrid reasoning, scores 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified while ranking 3rd overall and 2nd on agentic benchmarks.

  16. Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models

    cs.AI 2025-01 unverdicted novelty 3.0

    The paper surveys reinforced reasoning techniques for LLMs, covering automated data construction, learning-to-reason methods, and test-time scaling as steps toward Large Reasoning Models.