EQ-Bench Creative Writing v3

Measures: The model writes short fiction for 32 prompts x 3 iterations. An LLM judge (Claude Sonnet 4.6 since 2026-03-01; Sonnet 4 before) scores each piece on a rubric and also runs pairwise matchups. The matchups feed a Glicko-2 Elo that accounts for win margin, anchored at deepseek-r1 = 1500 and ministral-3b = 200.

Maintainer: Sam Paech (EQ-Bench)
Introduced: 2025
Status: Active
Saturation: Headroom
Typical provenance: Independent
In scoring set: Yes

Official leaderboard

Leaderboard

Show independent results only

Every result we hold, whoever ran it. Only independent and third-party rows count towards a score.

No results to show.