aiagents.org

Suspicion Agent

Research agent that plays imperfect-information games using theory-of-mind-aware GPT-4.

Visit site

Overview

Suspicion Agent is the implementation of a research project on playing imperfect-information games with theory-of-mind-aware GPT-4. The GitHub repository provides the project code for examining how language-model agents reason about other players and hidden information in game settings.

Key features

  • Imperfect-information game play
  • GPT-4-based agent
  • Theory-of-mind-aware reasoning
  • Research implementation

Use cases

Alternatives

  • AIliceOpen-source autonomous general-purpose agent for research, coding, system administration, and multi-step tasks.ai agent
  • Kimi AIAI assistant for long-context conversations, knowledge work, coding, and agentic tasks.ai agentPaid
  • MetaGPTMulti-agent framework that turns software requirements into plans, designs, tasks, and code.ai agent framework
  • AutomataOpen-source self-coding agent that generates code from project contextai agent
  • ChatDevMulti-agent software development system where LLM agents collaborate on coding workai agent
  • DemoGPTOpen-source tool for creating LLM agent applications from promptsai agent

Last updated 2026-09-16