Evaluation & BenchmarksAarohi Srivastava et al. (450 authors, 132 institutions)

Srivastava et al. release BIG-bench (204 tasks)

Released BIG-bench (204 tasks)

Why it matters

A community-crowdsourced benchmark. Concluded that performance improves with scale but stays low in absolute terms

Source — Primary sources first; secondary reports are labelled “Press”

Paperarxiv.orghttps://arxiv.org/abs/2206.04615

Added in Week 35, 2026 · See that week

Nearby on the same axis

Stanford CRFM releases HELMPaperLMSYS launches Chatbot ArenaOfficialMT-Bench and LLM-as-a-judge publishedPaperHugging Face explains the MMLU score mismatch on the Open LLM LeaderboardOfficial