Discover reinforcement learning AI agents: Browse our comprehensive directory of intelligent digital assistants and AI automation tools.

"reinforcement learning" tagged AI Agents | AI Agent Store

Kimi‑Dev (72B) is an open‑source coding language model by Moonshot AI, fine‑tuned via reinforcement learning to autonomously fix bugs and write tests, achieving state‑of‑the‑art SWE‑bench performance.

Kimi‑Dev

kimi-dev

Open‑source coding LLM optimized for automated debugging and issue resolution tasks.

MiniMax‑M1 is an open‑weight, hybrid Mixture‑of‑Experts large language model offering natively a 1 million token context window. It uses lightning attention and efficient RL training (CISPO) to excel at long-context reasoning, tool-use, coding, and mathematics. Released under Apache 2.0, it supports structured function calling and is deployable via vLLM, Transformers, or API.

MiniMax‑M1

minimax-m1

Open‑source large‑scale reasoning model with 1 million token context and hybrid Mixture‑of‑Experts architecture.

Agent Q is an advanced AI framework designed to improve the autonomy of AI agents in dynamic environments, such as web interfaces. It combines guided Monte Carlo Tree Search (MCTS), self-critique mechanisms, and reinforcement learning to enable agents to plan, execute, and adapt their actions effectively. This approach allows AI agents to handle complex, multi-step tasks with greater reliability and efficiency.

Agent Q

agent-q

AI framework enhancing autonomous agents' reasoning and learning in dynamic environments.

DeepSeek R1 is an open-source large language model (LLM) developed by the Chinese AI company DeepSeek. It is designed to excel in reasoning, math, coding, and problem-solving tasks. Released under the MIT license, it allows free access, modification, and commercialization, fostering collaboration and innovation. DeepSeek R1 has achieved remarkable benchmarks such as 97.3% on MATH-500 and 96.3% percentile on Codeforces, showcasing near-human performance in programming and logic-heavy tasks. Its unique training approach combines reinforcement learning (RL) and supervised fine-tuning (SFT), enabling it to learn autonomously while being cost-effective. DeepSeek R1 is a game-changer in democratizing AI development by making cutting-edge technology accessible to researchers, developers, and businesses worldwide.

DeepSeek R1

deepseek-r1

An open-source large language model excelling in reasoning, math, and coding tasks with MIT licensing for free use and modification.

OpenAI Operator is an AI agent developed to execute tasks on the web autonomously, mimicking human interactions with graphical user interfaces. Utilizing the 'Computer-Using Agent' (CUA) model, Operator combines GPT-4o's vision capabilities with reinforcement learning to navigate websites by typing, clicking, and scrolling without requiring custom API integrations. Initially available in the United States for ChatGPT Pro subscribers at $200 per month, Operator is designed to handle tasks such as online shopping, booking appointments, and managing emails. It self-corrects through reasoning and seeks user intervention for tasks requiring sensitive information. OpenAI has partnered with companies like DoorDash, Instacart, and Uber to enhance Operator's usability, with plans to expand access to Plus, Team, and Enterprise users in the future.

OpenAI Operator

openai-operator

An AI agent designed to autonomously perform web-based tasks by interacting with graphical user interfaces.

NVIDIA Eureka is an AI-driven agent designed to enhance robotic learning by autonomously generating reward algorithms that guide robots in mastering intricate skills. Utilizing OpenAI's GPT-4 large language model, Eureka creates software code that enables robots to learn through trial and error, significantly improving their performance in tasks such as pen spinning, opening drawers, and manipulating objects. This approach surpasses traditional human-authored programs, achieving over 50% performance improvement in more than 80% of evaluated tasks. Integrated with NVIDIA's Isaac Gym, a physics simulation platform, Eureka accelerates the training process by efficiently evaluating numerous reward candidates, facilitating rapid development of advanced robotic capabilities.

NVIDIA Eureka

nvidia-eureka

An AI agent developed by NVIDIA that leverages GPT-4 to autonomously train robots in complex tasks through reinforcement learning.

Yawning Titan is an abstract, graph-based cyber-security simulation environment designed to facilitate the training of intelligent agents for autonomous cyber operations. Developed by the Defence Science and Technology Laboratory (Dstl), it focuses on enabling defensive autonomous agents to counter probabilistic red (attacker) agents within simulated network environments. Built on OpenAI's Gym framework, Yawning Titan supports a wide range of reinforcement learning algorithms and offers flexible environment configurations, making it a valuable tool for research and development in cyber defense strategies.

Yawning Titan

yawning-titan

An open-source, graph-based cyber-security simulation environment for training intelligent agents in autonomous cyber defense operations.

LlamaGym is an open-source Python framework designed to simplify the fine-tuning of large language model (LLM) agents through online reinforcement learning. By providing a standardized environment similar to OpenAI's Gym, LlamaGym allows developers to efficiently train LLM-based agents by managing conversation context, episode batching, reward assignment, and proximal policy optimization (PPO) setup. This framework enables rapid experimentation with agent prompting and hyperparameters across various Gym environments, facilitating the development of more capable and responsive AI agents.

LlamaGym

llamagym

An open-source Python framework for fine-tuning large language model (LLM) agents using online reinforcement learning.

Explore AI Agents Tagged with "reinforcement learning"