P-Zero Research

How fast is AI's research taste improving?

We find that the experimental research taste of frontier models has doubled every ~3 months since December 2025. The best model, Opus 5.5, now exceeds our expert human baseline. Our human experts are experienced researchers, but most haven’t worked at a frontier lab.

Why measure research taste? In the AI Futures Model, it largely determines how quickly artificial superintelligence is reached once coding is fully automated.

Fable 5.2 is the best model we have tested at predicting experiment results. Its score is 0.36, in line with a steady improvement since GPT-4 Turbo.

Predicting results before running an experiment is a core part of research taste: a model that does it well can skip bad experiments and get more progress from the same compute.

We evaluated Kimi K3.1 Pro on TasteVal, and found that it has an experiment compute multiplier of 0.78x relative to our human experts, which puts it roughly 4.6 months behind the frontier and on par with Opus 4.8.

We fit a separate trend to leading open-weight models. On average, these models lag the frontier by about 4.2 months, and their doubling time is similar to that of frontier models.