The ARC Prize organization designs benchmarks which are specifically crafted to demonstrate tasks that humans complete easily, but are difficult for AIs like LLMs, "Reasoning" models, and Agentic frameworks.
ARC-AGI-3 is the first fully interactive benchmark in the ARC-AGI series. ARC-AGI-3 represents hundreds of original turn-based environments, each handcrafted by a team of human game designers. There are no instructions, no rules, and no stated goals. To succeed, an AI agent must explore each environment on its own, figure out how it works, discover what winning looks like, and carry what it learns forward across increasingly difficult levels.
Previous ARC-AGI benchmarks predicted and tracked major AI breakthroughs, from reasoning models to coding agents. ARC-AGI-3 points to what's next: the gap between AI that can follow instructions and AI that can genuinely explore, learn, and adapt in unfamiliar situations.
You can try the tasks yourself here: https://arcprize.org/arc-agi/3
Here is the current leaderboard for ARC-AGI 3, using state of the art models
- OpenAI GPT-5.4 High - 0.3% success rate at $5.2K
- Google Gemini 3.1 Pro - 0.2% success rate at $2.2K
- Anthropic Opus 4.6 Max - 0.2% success rate at $8.9K
- xAI Grok 4.20 Reasoning - 0.0% success rate $3.8K.
(Logarithmic cost on the horizontal axis. Note that the vertical scale goes from 0% to 3% in this graph. If human scores were included, they would be at 100%, at the cost of approximately $250.)
https://arcprize.org/leaderboard
Technical report: https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf
In order for an environment to be included in ARC-AGI-3, it needs to pass the minimum “easy for humans” threshold. Each environment was attempted by 10 people. Only environments that could be fully solved by at least two human participants (independently) were considered for inclusion in the public, semi-private and fully-private sets. Many environments were solved by six or more people. As a reminder, an environment is considered solved only if the test taker was able to complete all levels, upon seeing the environment for the very first time. As such, all ARC-AGI-3 environments are verified to be 100% solvable by humans with no prior task-specific training
You can really only judge fairness of the score if you understand the scoring criteria. It is a relative score where the baseline is 100% for humans -- i.e. A task was only included in the challenge if at least two people in the panel of humans were able to solve it completely, and their action count is a measure of efficiency. This is the baseline used as a point of comparison.
From the Technical Report:
So the humans "scored 100%" because that is the baseline by definition, and the AIs are evaluated at how close they got to human correctness and efficiency. So a score of 0.26% is 1/0.0026 ~= 385 times less efficient (and correct) compared to humans.
Right. Thank you for this explanation, the percentages seemed out of context. So, the LLM was able to complete some levels?
If you look at the list of tasks,, you can see how the 4 frontier models did. Some of them did complete one or two levels of one or two tasks. None of them completed a whole task. Some of the reasoning logs are funny in the replays.
Yes, the LLMs received credit for each level even if they didn't complete the entire environment.
They have some replays of tasks on their website: https://arcprize.org/tasks
Here's one where the human completed all 9 levels in 1458 actions, but the LLM completed only one level in 24 actions, then struggled for 190 actions until it timed-out, I guess. The LLM scored 2.8% because of the weighted average, I think. I didn't take the time to all do the math, and I'm not sure if the replay action count is accurate, but it gives you an idea.
Human: https://arcprize.org/replay/0d461c1c-21e5-4dc8-b263-9922332a6485
LLM: https://arcprize.org/replay/cc821983-3975-4ae4-a70b-e031f6807bb0