Benchmark
InnovationEval
An evaluation benchmark introduced by Epoch AI to assess whether AI models can independently complete a full research cycle—from ideation and implementation to experimentation—producing novel algorithmic innovations comparable to those produced by human researchers.
Timeline (last 90 days)
Oct 12
Epoch AI's InnovationEval is used to evaluate whether a model can independently produce new algorithms—from conception and implementation to experimentation—that are comparable to techniques subsequently discovered by humans.
Unverified50%