Category

Recursive Game Creator: An Agentic Product-Level Experience-Oriented Game Harness

Iteratively refines agent-generated games using a closed loop of Designer, Builder, Coding‑Native Player, and Experience‑Oriented Reviewer — collects programmatic gameplay trajectories and trajectory+visual evaluations to evolve prototypes into player-focused, product-level games; shows measurable gains on GameCraft-Bench and GameASG-Bench.

nanoMuse: An Open-Source Personal Agent for Every Device You Own

Runs an open-source personal software agent across a user's devices to perform actions, keep readable provenance-backed memories, and coordinate via a self-hostable relay. Ships minimal end-to-end clients (mobile, desktop, web) under GPL-3.0; model-agnostic and designed for inspectability and device

Learn2Play Bench: How Well Do LLM Agents Learn from Experience in Unfamiliar Environments?

Measures how effectively LLM agents learn from interaction by playing 20 text-based games with novel or counterintuitive hidden rules, providing deterministic feedback, episode-wise scoring, and controlled variations to test retention and transfer.

AgentGarten: Code Worlds for Evolving Agents

Creates programmable, real-time interactive code-based environments by separating deterministic simulator state from a shared neural video renderer so agents perceive, interact, and iteratively evolve via distilled playbooks; introduces Adversarial Forcing to distill a geometry-conditioned renderer for responsive visual feedback.

SuperNav: An Agentic Navigation System for Any Task in Any Scene

Lets a pretrained multimodal LLM interpret navigation requests and orchestrate motion via tool calls for generalist robot navigation across unfamiliar scenes. Key features: an agent harness with Navigation Skills, a unified visual-point interface, task-progress tracking, and tool-based motion execution without navigation-specific fine-tuning.

ProgramGym: Scaling Training Environments for Long-Horizon Program Reproduction Agents

Converts real repositories into verifiable, sandboxed training environments with black-box tests to train coding agents on long-horizon program reproduction. Provides Program-Env for automated environment construction and Program-Agent for a role-isolated development workflow, yielding 1,648 C/C++/Go/Rust tasks and improving benchmark performance.

Scaling Agentic Data for Long-Horizon Terminal Intelligence

Proposes TerminalHorizon, an automated engine that synthesizes long-horizon task environments and agentic trajectories that preserve exploration, evidence accumulation, and iterative improvement; uses ~1.5K environments and ~2.8K trajectories to fine-tune Qwen3.5-27B and improves performance on multiple benchmarks.

AgentSWE: Can Coding Agents Build the Agent You Actually Want?

Measures whether coding agents can build agents that truly satisfy natural-language requirements via AgentSWE: a 25-task benchmark spanning Creation, Editing, and Optimization. Builders develop against few illustrative cases; acceptance runs frozen submissions on held-out cases and requirement-derived rubrics, exposing overfitting, hollow runtimes, and unsupported completion claims.

Into the Unknown: Benchmarking AI agents on open scientific research problems

Evaluates whether AI agents can autonomously carry out and honestly report follow-up scientific research by giving each agent a published paper plus a scientist-defined follow-up task and rubrics; measures integrity, completeness, and sufficiency across 15 open research tasks.

Jeva: System One Models from First Principles

Presents Jeva, a decision-native architecture that encodes requests and each candidate separately to produce permutation-invariant, option-scalable probability distributions with explicit null detection and reusable option representations for low-latency decision tasks.

Self-Evolving Agent Harness

Defines a feedback-driven lifecycle that lets an agent's harness (prompts, memory, workflows, tools, control logic) iteratively self-revise across runs while keeping model weights fixed. Distinguishes interpreted vs executable updates, frames search as evolutionary, and gives a fresh-validation admission rule.

Urbex: Agentic Spatial Grounding for City-Scale 3D Scenes

Reframes city-scale 3D visual grounding as active spatial search: a VLM agent interacts with a multi-view BEV environment to query landmarks, zoom, render oblique views, and commit world-aligned bounding boxes. Trains a tool-use policy with reinforcement learning on CityRefer, improving Hit@1, IoU, and inference speed over ranking baselines.