The best AI for writing romance
Live data · ranked by Romance · measured as of July 23, 2026
The difference between a couple a reader roots for and two names told to be in love is charge — the tension of attraction, the pull between characters, written rather than reported. This ranking scores the frontier models on romance short of the explicit, and the leader is the model built for exactly this.
Key takeaways
- Models are ranked by measured romance — attraction, tension, chemistry short of the explicit — on real novel prose.
- Fable 5, Anthropic’s storytelling model, leads romance, with GPT-5.6 Sol right alongside.
- Romance and explicit content are separate axes — a model can write the best romance and still stop short of the explicit (see the explicit-scenes ranking).
- The ranking is live and updates as more prose is measured.
The ranking
Ranked by Romance
Ranked highest-to-lowest by measured romance writing. The score is each model’s confidence-adjusted value on the romance axis, 0–100.
- GPT-5.6 SolOpenAI83
Right alongside Fable on romance, with far stronger action and dialogue — a good pick for romance woven into a plot-driven book.
- Fable 5Anthropic83
Leads romance. Anthropic’s storytelling model writes the pull between characters rather than reporting it — the strongest measured choice for romance-forward fiction.
- GPT-5.6 TerraOpenAI81
Strong romance at lower cost — Sol’s shape one step down.
- Claude Opus 4.8Anthropic · Novelmint Lead80
- Claude Sonnet 5Anthropic75
- Gemini 3.1 ProGoogle74
- Gemini 3.5 FlashGoogle73
- Grok 4.5xAI73
- Grok 4.3xAI73
- GPT-4.1OpenAI70
- Claude Haiku 4.5Anthropic68
How this ranking is made
Scored on charge, not heat
Each model writes the same briefed romance beats, and a blind judge panel rates the charge between characters — tension, chemistry, the pull — short of explicit content, without knowing the author. Thin samples are shrunk toward a neutral baseline. This page orders the models by that one axis. Explicit content is a separate axis with its own ranking.
See the full nine-axis benchmark and methodologyQuestions
Frequently asked
- Which AI writes romance best?
- Fable 5, Anthropic’s storytelling model, leads measured romance in the Novelmint benchmark, with GPT-5.6 Sol right alongside. The live ranking shows the current order.
- Is the best romance model also best for explicit scenes?
- No — romance (charge and chemistry) and explicit content are separate axes. Fable leads romance but stops short of the explicit; for explicit scenes a permissive model leads. See the explicit-scenes ranking.
- How is romance measured?
- A blind judge panel rates romance beats on real novel prose for the charge between characters — tension, chemistry — short of the explicit, with low-sample scores shrunk toward a neutral baseline.
What this page does not claim
- This ranks measured romance (short of explicit) on fiction beats only.
- Romance and explicit content are scored separately; this ranking is not about explicit scenes.
- Model names are trademarks of their respective owners; this is an independent measurement.
関連
AI fiction model benchmark
How the frontier models actually write fiction — measured across nine axes.
Best AI for action scenes
Which AI keeps a fight or chase clear and grounded instead of vague. Ranked on physicality.
Best AI for conflict scenes
Which AI keeps an argument or standoff taut and rising. Ranked on measured conflict.
Best AI for emotional scenes
Which AI makes a feeling-led scene land instead of just describing it. Ranked on emotion.
Give your love story the model with the chemistry.
Novelmint routes romance beats to the model that wins them. Your first chapter is free.