Hacker News
new
top
best
ask
show
job
Wandr Benchmark: Evaluating Research Agents That Must Search Wide and Deep
(
research.perplexity.ai
)
1 point
by
tagawa
8 hours ago
1 comment
tagawa
8 hours ago
Repo with benchmark tasks, evaluation harness, tech report:
https://github.com/perplexityai/wandr