Importantly: this is a snapshot of mid-2025 novice and LLM performance.
Results could change as new LLMs become more capable, easier to use in the lab, and as average elicitation skill improves.
As models evolve, we aim to continue tracking how people use frontier AI in biology