Fiction Eval leaderboard · comparison

Gemini 3.1 Pro vs Llama 4 Maverick for fiction writing

Judged blind on fresh genre scenes (LitRPG, cozy mystery, romance, thriller, xianxia…), July 2026 — three tries per model per scene, passages shown in both orders so position can't sway it, graded by a third AI that isn't either of them.

Head-to-head

Gemini 3.1 Pro won the direct matchup 28–0 (with 0 ties) across every prompt and sample.

#ModelScoreRecordSlop
5Gemini 3.1 Pro1471 (likely 1423–1519)78–104–1313.45
8Llama 4 Maverick686 (likely 627–741)0–206–122.92

The score is each model's overall strength across the whole 8-model field (like a chess rating); the range next to it is where the true score most likely sits. Slop is a separate count of AI tells (lower is cleaner) — it isn't part of the ranking.

The same scene, both models

Cozy mystery — opening of the "cozy cold open" scene. Full passages for all models are on the board.

Gemini 3.1 Pro:

The sea breeze off Pelican Bay usually brought the scent of brine and drying kelp, but by six o’clock every morning, it was thoroughly overpowered by butter. Beatrice Gable, proprietor of The Salted Crumb, had a heavy hand with the French butter, which was exactly why the line for her croissants typically wrapped around the block. But this morning—the morning of the Pelican Bay Autumn Harvest Festival—my bookshop smelled only of old paper and the sea. I checked the brass clock behind my counter. Six-fifteen. Bea was late with my morning Americano and bear claw. In the three years since I’d opened *Tomes & Tides* next door, Bea had never been late. She operated with the precision of a Swiss watch, assuming the watch was coated in flour and prone to singing Patsy Cline off-key. Dropping my feather duster, I pushed through the adjoining door that connected our stockrooms. "Bea?" The bakery kitchen was silent. The massive industrial ovens, usually radiating a comforting, bread-scented heat, were dark and stone-cold. That …

Llama 4 Maverick:

I stepped out of my bookshop, the scent of saltwater and the chatter of the morning crowd greeting me like an old friend. The sun was shining, the sea was calm, and the smell of sugar and spices wafted from next door, where The Sweet Spot bakery was getting ready for the harvest festival. Agatha's pastry cases would be filled to bursting with her famous apple cider donuts, and her display windows would be a riot of colorful leaves and pumpkins. But as I pushed open the bakery door, the warm glow of the morning was snuffed out. The police were already there, cordoning off the area with yellow tape. I spotted Chief Thompson's stern face, and my heart sank. Something was very wrong. "What's going on?" I asked the officer stationed at the door, trying to peer past him into the bakery. "Agatha's been found dead in her walk-in," he said, his voice low and serious. I felt a pang of shock, followed by a wave of sadness. Agatha …

More

Gemini 3.1 Pro full review · Llama 4 Maverick full review · the July 2026 report · all comparisons

How it works: every pair of models is judged blind on the same scene, with the passages' order flipped so being shown first can't sway it, by GPT-5.4 — a family that isn't on the board, so nobody scores their own side. Each score carries a likely range; overlapping ranges are called a tie. Slop is scored separately by a fixed checklist, not an AI. Full board, every prompt, and the FAQ: thebookfactoryai.com/board. Get each new board by email on the model-drop list.

Writing a book of your own? Book Factory runs the same craft checks on full manuscripts.