Tufa Labs Wins $25K Beating Frontier AI on the World's Hardest Benchmark
ARC Prize awards its first Milestone #1 prizes for ARC-AGI-3, with Tufa Labs, Reki, and Md Boktiar Mahbub Murad open-sourcing the top-scoring agents on the world's hardest agentic benchmark.
- Milestone #1 winners announced: Tufa Labs (1st, $25K), Reki (2nd, $10K), and Md Boktiar Mahbub Murad (3rd, $2.5K) open-sourced top ARC-AGI-3 agents.
- Scores are tiny but meaningful: Tufa Labs' winning "duck harness" scored 1.03 on Kaggle; frontier LLMs like GPT-5.4 and Gemini score below 0.5%.
- ARC-AGI-3 is a new kind of benchmark: Agents must explore interactive game environments with no instructions, rules, or stated goals — humans score 100%.
- RL beats LLMs: Tufa Labs' StochasticGoose uses a CNN + RL approach to predict action-causing frame changes, outperforming all frontier language models.
- $700K grand prize still unclaimed: The ARC-AGI-3 track offers $850K total; the grand prize requires 100% accuracy and rolls over to 2027 if unclaimed.
- All three notebooks are now public on Kaggle, giving the community working baselines ahead of the September 30 Milestone #2 deadline.
The hardest AI benchmark alive just handed out its first prize money. ARC Prize 2026 has awarded its Milestone #1 prizes for the ARC-AGI-3 competition, recognizing the top three teams who open-sourced their agents on the Kaggle leaderboard by the June 30 deadline. The winners: Tufa Labs in first, Reki in second, and Md Boktiar Mahbub Murad in third. All three notebooks are now public.
A benchmark that breaks everything frontier AI has
ARC-AGI-3 is a fundamentally different kind of benchmark. It comprises hundreds of original turn-based environments, each handcrafted by human game designers, with 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.
This marks the first major format change since ARC-AGI was introduced in 2019. While the first two versions challenged static reasoning, the third version is designed to challenge interactive reasoning and requires new AI capabilities: exploration, planning, memory, goal acquisition, and alignment.
The gap between human and AI performance is staggering. Humans can solve 100% of the environments, in contrast to frontier AI systems which, as of March 2026, score below 1%. To put that in context: by early 2026, Gemini 3.1 Pro was scoring 98% on ARC-AGI-1. That benchmark was effectively solved. ARC-AGI-3 reset the clock entirely.
The scoring system is also novel. Due to the operational intensity of running a full evaluation using high-reasoning frontier model APIs, there is a hard cutoff of 5x human performance per level. If a human takes 10 actions to beat a certain level on average, the AI agent is cut off after 50 actions. Efficiency is the metric, not just completion.
The milestone prize structure
Participants who open-source their solutions by the milestone deadlines are eligible for milestone prize money. Milestone #1 (June 30, 2026): 1st place receives $25K, 2nd receives $10K, and 3rd receives $2.5K. A second milestone checkpoint follows on September 30 with the same payout structure.