Augur maps where a model breaks — where its knowledge ends and its reasoning fails — and turns that map into a benchmark that cannot be memorized and the exact training data to move the boundary.

Augur maps where a model breaks — where its knowledge ends and its reasoning fails — and turns that map into a benchmark that cannot be memorized and the exact training data to move the boundary.
RSI is the fastest path to superintelligence, and it is already here. We make the term precise, map the approaches being called RSI today, and explain why improving the system rather than the weights lets the loop compound quickly, cheaply, and under tighter control.
With zero human intervention, Poetiq's Metasystem built harnesses that set new state-of-the-art results across six diverse benchmarks — often using a generation-older model than the previous SOTA holder.
How Poetiq's Meta-System topped a coding benchmark without fine-tuning or privileged model access.
Following our milestone on ARC-AGI, we used the Poetiq Meta-System to achieve new state-of-the-art (SOTA) results on Humanity’s Last Exam (HLE) and SimpleQA. We also improve performance on multiple frontier models.
We are thrilled to announce that Poetiq has raised $45.8M in Seed funding. This round was supported by a coalition of world-class partners.
We finally had a moment to run our system with GPT-5.2 X-High on ARC-AGI-2!
We are proud to confirm that our system has officially outperformed existing methods, establishing a new state-of-the-art by a significant margin.
Poetiq is proud to announce a major milestone in AI reasoning. We have established a new state-of-the-art (SOTA) on the ARC-AGI-1 & 2 benchmarks, significantly advancing both the performance and the efficiency of the current AI systems.