Orak
Nemotron 3 Nano 4B 22.9 Fading LLM agents play 12 popular video games across most genres through an MCP interface. Scores come from in-game progress or outcomes per title.
The name is the Korean word for game. The repo also releases gameplay data for fine-tuning.
In the authors’ words
Large Language Model (LLM) agents are reshaping the game industry, by enabling more intelligent and human-preferable characters. Yet, current game benchmarks fall short of practical needs: they lack evaluations of diverse LLM capabilities across various game genres, studies of agentic modules crucial for complex gameplay, and fine-tuning datasets to adapt pre-trained LLMs into gaming agents. To fill these gaps, we present Orak, a benchmark for training and evaluating LLM agents across 12 popular video games spanning all major genres. Using a plug-and-play interface built on Model Context Protocol (MCP), Orak supports systematic and reproducible studies of agentic modules in varied game scenarios. We further release a fine-tuning dataset of expert LLM gameplay trajectories covering multiple genres, turning general LLMs into effective game agents. Orak offers a united evaluation framework, including game leaderboards, LLM battle arenas, and \fixablation studies of input modality, agentic strategies, and fine-tuning effects, establishing a foundation towards versatile gaming agents. Code and datasets are available at https://github.com/krafton-ai/Orak and https://huggingface.co/datasets/KRAFTON/Orak.
- Level
- Everyday Most adults, without training
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Agent task Many steps in a terminal, browser, desktop, app, game or robot.
Nemotron 3 Nano 4B leads with 22.9
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Nemotron 3 Nano 4B | 22.9 | Mar 2026 | Model card |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Labs headline it less often now, last in Mar 2026
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 16 Mar 2026 | Nemotron 3 Nano 4B | 22.9 | Orak (reasoning off) |
More in Games and long tasks
- ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6 Astra 100%
- MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks.
Seed2.0 Pro 49% - GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state.
Gemini 3 Flash 41.9% - Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%
- TextQuests Only in system cards Agents play 25 classic interactive fiction games. Measures game progress and harmful actions per 100 steps.
Muse Spark 2.40 - Anthropic large agent teams tasks Only in system cards Internal 24-hour tasks for teams of 1-100 agents: build a company knowledge base, or formalize a theorem in Lean.