Benchmark Almanac

Flash Fiction Stylistic Diversity

GPT-5 0.22 leads since September 2025
Not in model cards

Models write flash fiction from shared prompts. A reviewing model describes each story, and weighted Gower distance measures stylistic diversity within a writer’s work.

The separate diversity leaderboard reports 29 models from 15,347 stories.

Made by
Lech Mazur
Released
September 2025
Kind
Taste and style
Level
School
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5.2 #51 on the capability index
Last model added
December 2025
Scores refreshed
November 21st, 2026
28
Entries on its own leaderboard, 23 models we track

GPT-5 leads with 0.22 among models released by December 2025 GPT-5 leads with 0.22 among models released by December 2025

Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in December 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-5 medium 0.22
  2. 2 GPT-5.2 medium 0.21
  3. 3 GPT-5 Pro 0.21
  4. 4 GPT-5.1 medium 0.21
  5. 5 Kimi K2 Thinking 0.21
  6. 6 GLM-4.5 0.21
  7. 7 Grok 4.1 Fast reasoning 0.20
  8. 8 DeepSeek-V3.2-Exp 0.20
  9. 9 Claude Opus 4.1 no reasoning 0.20
  10. 10 Kimi K2 Sep 2025 0.20
  11. 11 GLM-4.6 0.20
  12. 12 Qwen 3 235B A22B 25-07 Think 0.20
  13. 13 Claude Sonnet 4.5 Thinking 16K 0.20
  14. 14 o3-pro medium 0.20
  15. 15 Claude Sonnet 4.5 no reasoning 0.20
  16. 16 Mistral Large 3 0.20
  17. 17 Baidu Ernie 4.5 300B A47B 0.19
  18. 18 Grok 4 0.19
  19. 19 Gemini 2.5 Pro Jun 2025 0.19
  20. 20 Gemini 3 Pro 0.19
  21. 21 DeepSeek V3.1 Non-Think 0.19
  22. 22 DeepSeek V3.1 Reasoner 0.19
  23. 23 Mistral Medium 3.1 0.19
  24. 24 Claude Opus 4.5 no reasoning 0.19
  25. 25 Llama 4 Maverick 0.18
  26. 26 Qwen3-Max-Preview 0.18
  27. 27 Cohere Command A 0.18
  28. 28 gpt-oss-120b 0.17

Source: Official Flash Fiction Stylistic Diversity leaderboard

#EntryScoreSettingDateSource
1 GPT-5 0.22 medium Aug 2025 Official leaderboard
2 GPT-5.2 0.21 medium Dec 2025 Official leaderboard
3 GPT-5 Pro 0.21 Oct 2025 Official leaderboard
4 GPT-5.1 0.21 medium Nov 2025 Official leaderboard
5 Kimi K2 Thinking 0.21 Nov 2025 Official leaderboard
6 GLM-4.5 0.21 Aug 2025 Official leaderboard
7 Grok 4.1 Fast 0.20 reasoning Nov 2025 Official leaderboard
8 DeepSeek-V3.2-Exp 0.20 Sep 2025 Official leaderboard
9 Claude Opus 4.1 0.20 no reasoning Aug 2025 Official leaderboard
10 Kimi K2 Sep 2025 0.20 Sep 2025 Official leaderboard
11 GLM-4.6 0.20 Sep 2025 Official leaderboard
12 Qwen 3 235B A22B 25-07 Think 0.20 Official leaderboard
13 Claude Sonnet 4.5 Thinking 16K 0.20 Official leaderboard
14 o3-pro 0.20 medium Jun 2025 Official leaderboard
15 Claude Sonnet 4.5 0.20 no reasoning Sep 2025 Official leaderboard
16 Mistral Large 3 0.20 Dec 2025 Official leaderboard
17 Baidu Ernie 4.5 300B A47B 0.19 Official leaderboard
18 Grok 4 0.19 Jul 2025 Official leaderboard
19 Gemini 2.5 Pro Jun 2025 0.19 Jun 2025 Official leaderboard
20 Gemini 3 Pro 0.19 Nov 2025 Official leaderboard
21 DeepSeek V3.1 Non-Think 0.19 Official leaderboard
22 DeepSeek V3.1 Reasoner 0.19 Official leaderboard
23 Mistral Medium 3.1 0.19 Aug 2025 Official leaderboard
24 Claude Opus 4.5 0.19 no reasoning Nov 2025 Official leaderboard
25 Llama 4 Maverick 0.18 Apr 2025 Official leaderboard
26 Qwen3-Max-Preview 0.18 Sep 2025 Official leaderboard
27 Cohere Command A 0.18 Mar 2025 Official leaderboard
28 gpt-oss-120b 0.17 Aug 2025 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5 0.22 medium Aug 2025 Official leaderboard
2 GPT-5.2 0.21 medium Dec 2025 Official leaderboard
3 GPT-5 Pro 0.21 Oct 2025 Official leaderboard
4 GPT-5.1 0.21 medium Nov 2025 Official leaderboard
5 Kimi K2 Thinking 0.21 Nov 2025 Official leaderboard
6 GLM-4.5 0.21 Aug 2025 Official leaderboard
7 Grok 4.1 Fast 0.20 reasoning Nov 2025 Official leaderboard
8 DeepSeek-V3.2-Exp 0.20 Sep 2025 Official leaderboard
9 Claude Opus 4.1 0.20 no reasoning Aug 2025 Official leaderboard
10 Kimi K2 Sep 2025 0.20 Sep 2025 Official leaderboard
11 GLM-4.6 0.20 Sep 2025 Official leaderboard
12 o3-pro 0.20 medium Jun 2025 Official leaderboard
13 Claude Sonnet 4.5 0.20 no reasoning Sep 2025 Official leaderboard
14 Mistral Large 3 0.20 Dec 2025 Official leaderboard
15 Grok 4 0.19 Jul 2025 Official leaderboard
16 Gemini 2.5 Pro Jun 2025 0.19 Jun 2025 Official leaderboard
17 Gemini 3 Pro 0.19 Nov 2025 Official leaderboard
18 Mistral Medium 3.1 0.19 Aug 2025 Official leaderboard
19 Claude Opus 4.5 0.19 no reasoning Nov 2025 Official leaderboard
20 Llama 4 Maverick 0.18 Apr 2025 Official leaderboard
21 Qwen3-Max-Preview 0.18 Sep 2025 Official leaderboard
22 Cohere Command A 0.18 Mar 2025 Official leaderboard
23 gpt-oss-120b 0.17 Aug 2025 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
Own leaderboard Not tested since Dec 2025
GPT-5 0.22

Source: Official Flash Fiction Stylistic Diversity leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
0.18 0.20 0.22 2025 2026 0.18 0.20 0.22 2025 2026

Source: Official Flash Fiction Stylistic Diversity leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Writing and taste

All 65 in Writing and taste
  • Design Bench (Moonshot) In use Models make graphic designs from direct prompts or few-shot reference images. An automated grader rates the results as Elo scores. Kimi K3.1 Pro 957
  • WritingBench Fading Models complete 1,000 real-world writing requests across 6 domains and 100 subdomains. A judge model scores each reply on five query-specific criteria. GPT-5 83.9
  • Creative Writing (MiniMax in-house) Only in system cards Models write creative pieces for real Hailuo AI user requests. Automatic assessments score the writing. MiniMax-Text-01 81.3%
  • DeepSeek Chinese functional writing evaluation Only in system cards Models answer 3,170 everyday Chinese writing prompts such as reports, emails and speeches. Pairwise judging against a baseline model gives a win rate. DeepSeek-V4-Pro 62.6%
  • Slide generation human evaluation (Z.ai) Only in system cards People compare HTML slide decks from two models on content, layout and visual quality. Metric is overall win rate. GLM-5 67.5%
  • Task-dependent output homogenization evaluation Only in system cards Models answer prompts several times. A judge scores functional diversity where it is wanted and consistency where one answer is right. Muse Spark 0.21

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 11 months ago, in December 2025, for GPT-5.2. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (28 entries).
  • Pass Models scored 23 models we track, plus 5 other entries on its leaderboard.
  • Warning Newest score The newest came 11 months ago, in December 2025, for GPT-5.2.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.2 (#51).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 24 rows for models we track and 5 other entries at the last read.
  • Pass Last read November 21st, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check