Wandr Benchmark: Evaluating Research Agents That Must Search Wide and Deep

1 comments

• tagawa an hour ago

  Repo with benchmark tasks, evaluation harness, tech report:

  https://github.com/perplexityai/wandr