Writing quality can't be reduced to one test, so this ranking triangulates two public boards that measure it differently: LMArena's creative-writing category, built from blind human votes, and EQ-Bench's Creative Writing leaderboard, judged on narrative quality and emotional depth. The boards openly disagree about OpenAI's latest model — we follow the human voters first, and say so where the order comes from.
Anthropic's thoughtful AI assistant for writing, coding, and reasoning.
Claude Opus leads LMArena's human-voted creative-writing board and sits at the top of EQ-Bench's longform table — the writers' consensus pick for prose that needs the least editing.
The world's most popular AI chatbot.
GPT-6 Astra leads EQ-Bench's AI-judged Creative Writing board, even though LMArena's human voters rank it lower — genuinely contested, and excellent at structure and marketing copy.
Google's AI assistant woven into Search, Gmail, and Android.
A consistent top-five finisher in LMArena's category votes, with a very large context window that keeps a whole manuscript in view.
xAI's witty AI assistant
Frontier-tier prose with a looser, more distinctive voice than its rivals — a regular in the top group of LMArena's text rankings.
AI content platform for businesses.
The specialist platform for marketing teams: brand voice, campaigns and content workflows built around frontier models.
AI copywriting for marketing teams.
Go-to-market copy at scale — sales pages, ads and outbound workflows rather than long-form prose.
Alibaba's capable AI chat assistant
Strong open-weight writing, especially multilingual drafting, at a fraction of frontier-model prices.
Mistral's fast AI assistant
Fast, fluent European alternative for everyday drafting and rewriting, with a lighter touch than the top three.
Ranking basis: LMArena Creative Writing leaderboard and EQ-Bench Creative Writing. Order reflects those two public boards as of the check date: human preference (LMArena) weighted first, judged narrative quality (EQ-Bench) second. We publish ranks, not scores. Data checked October 2026. Rankings are editorial, ordered from the public sources named here; Tools Hub does not accept paid placements on this list.
By blind human votes on LMArena's creative-writing board, Anthropic's Claude leads, and it also tops EQ-Bench's longform writing table. OpenAI's GPT leads EQ-Bench's judged creative-writing board instead — if you write fiction, Claude is the consensus first try; if you write structured marketing copy, GPT is at least its equal.
LMArena shows voters two anonymous answers to a writing prompt and counts which they prefer. EQ-Bench scores generated stories and essays for narrative quality and emotional depth. Neither is perfect — voters reward fluency, judges reward structure — which is why this page cross-checks both.
If you write in a team with a brand voice, yes: Jasper and Copy.ai (both on this list) wrap frontier models in templates, voice controls and approval workflows that a raw chatbot lacks. Solo writers usually get further with a top model and a good brief.