WildTrace Benchmark

What is WildTrace?

WildTrace evaluates long-context reasoning over naturally occurring long-form sources, focusing on integrating evidence dispersed across distant passages. It…

Released

2026-07-10

Evaluates

General AI, Language & Knowledge, Long Context, Reasoning

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2607.09328

WildTrace paper

• Paper https://arxiv.org/abs/2607.09328

Which sources cite WildTrace?

• WildTrace

Related General AI, Language & Knowledge, Long Context, Reasoning benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A-OKVQA
• A3S-Bench
• AARRI-Bench
• ABC-Bench