Benchmark Almanac

Fiction Bench (Story Hacker)

GPT-6.2 Astra 73.6 leads since November 2026
Not in model cards

Models write three chapters for each of 20 genres. The overall score weights blind reader preferences at 60 percent and model judges’ craft scores at 40 percent.

The organizer’s leaderboard has scores for 42 models that completed all 20 stories. Human ratings with fewer than 20 votes are provisional.

Made by
Story Hacker AI
Released
October 2026
Kind
Creative writing
Level
University
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Measures
Preference
Language
English
In and out
Text → Text
Interaction
Single turn
Items from
Live play or votes
Refresh
One fixed set
Run by
A host
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
42
Entries on its own leaderboard, 36 models we track

GPT-6.2 Astra leads with 73.6 GPT-6.2 Astra leads with 73.6

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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.

  1. 1 GPT-6.2 Astra Provisional human rating 73.6
  2. 2 GPT-6.2 Sol Provisional human rating 72.6
  3. 3 GPT-6.1 Terra 68.3
  4. 4 Claude Opus 5.5 66.3
  5. 5 Gemini 4 Argon 65.1
  6. 6 Gemini 4 Carbon 64.4
  7. 7 Muse Wonder 64.0
  8. 8 Claude Fable 5.1 63.6
  9. 9 Claude Sonnet 5.5 63.0
  10. 10 GPT-6 Astra 62.7
  11. 11 Kimi K3.1 Pro 62.5
  12. 12 GPT-6 Sol 62.3
  13. 13 GPT-6 Astra Pro 62.2
  14. 14 Claude Fable 5.2 62.1
  15. 15 GPT-6 Luna 61.3
  16. 16 Muse Spark 1.3 61.0
  17. 17 Claude Opus 4.8 60.7
  18. 18 Gemini 3.7 Flash 60.3
  19. 19 GPT-6.1 Sol 60.0
  20. 20 Gemini 3.8 Flash 59.7
  21. 21 aion-labs/aion-3.5 59.6
  22. 22 Kimi K3 58.5
  23. 23 aion-labs/aion-3.0 57.3
  24. 24 aion-labs/aion-3.5-mini 56.9
  25. 25 DeepSeek V4.1 Flash 1017 56.8
  26. 26 Qwen 3.8 Max prime 55.9
  27. 27 GLM-5.3-Flash 55.2
  28. 28 MiMo-V2.6 Flash 55.2
  29. 29 Muse Ember 54.9
  30. 30 Claude Haiku 5.5 54.8
  31. 31 MiniMax-M3 54.5
  32. 32 Grok 4.7 54.1
  33. 33 MiMo-V2.6 Pro 53.9
  34. 34 GLM-5.3 52.3
  35. 35 Gemini 3.5 Flash-Lite 52.2
  36. 36 DeepSeek V4 Pro 0813 51.6
  37. 37 DeepSeek V4.1 Flash 51.6
  38. 38 aion-labs/aion-3.0-mini 51.2
  39. 39 Muse Glimmer 50.0
  40. 40 aion-labs/aion-2.0 49.6
  41. 41 Qwen3.8 Flash 46.1
  42. 42 mistralai/mistral-large-4-0 43.2

Source: Official Fiction Bench (Story Hacker) leaderboard

Source: Official Fiction Bench (Story Hacker) leaderboard

#EntryScoreSettingReleasedRun onSource
1 GPT-6.2 Astra 73.6 Provisional human rating Nov 2026 Nov 2026 Official leaderboard
2 GPT-6.2 Sol 72.6 Provisional human rating Nov 2026 Nov 2026 Official leaderboard
3 GPT-6.1 Terra 68.3 Oct 2026 Oct 2026 Official leaderboard
4 Claude Opus 5.5 66.3 Sep 2026 Oct 2026 Official leaderboard
5 Gemini 4 Argon 65.1 Sep 2026 Oct 2026 Official leaderboard
6 Gemini 4 Carbon 64.4 Oct 2026 Nov 2026 Official leaderboard
7 Muse Wonder 64.0 Oct 2026 Oct 2026 Official leaderboard
8 Claude Fable 5.1 63.6 Sep 2026 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 63.0 Sep 2026 Oct 2026 Official leaderboard
10 GPT-6 Astra 62.7 Sep 2026 Oct 2026 Official leaderboard
11 Kimi K3.1 Pro 62.5 Oct 2026 Oct 2026 Official leaderboard
12 GPT-6 Sol 62.3 Sep 2026 Oct 2026 Official leaderboard
13 GPT-6 Astra Pro 62.2 Sep 2026 Oct 2026 Official leaderboard
14 Claude Fable 5.2 62.1 Nov 2026 Nov 2026 Official leaderboard
15 GPT-6 Luna 61.3 Sep 2026 Oct 2026 Official leaderboard
16 Muse Spark 1.3 61.0 Sep 2026 Oct 2026 Official leaderboard
17 Claude Opus 4.8 60.7 May 2026 Oct 2026 Official leaderboard
18 Gemini 3.7 Flash 60.3 Aug 2026 Oct 2026 Official leaderboard
19 GPT-6.1 Sol 60.0 Sep 2026 Oct 2026 Official leaderboard
20 Gemini 3.8 Flash 59.7 Sep 2026 Oct 2026 Official leaderboard
21 aion-labs/aion-3.5 59.6 Oct 2026 Official leaderboard
22 Kimi K3 58.5 Jul 2026 Oct 2026 Official leaderboard
23 aion-labs/aion-3.0 57.3 Oct 2026 Official leaderboard
24 aion-labs/aion-3.5-mini 56.9 Oct 2026 Official leaderboard
25 DeepSeek V4.1 Flash 1017 56.8 Oct 2026 Oct 2026 Official leaderboard
26 Qwen 3.8 Max 55.9 prime Aug 2026 Oct 2026 Official leaderboard
27 GLM-5.3-Flash 55.2 Aug 2026 Oct 2026 Official leaderboard
28 MiMo-V2.6 Flash 55.2 Sep 2026 Oct 2026 Official leaderboard
29 Muse Ember 54.9 Nov 2026 Nov 2026 Official leaderboard
30 Claude Haiku 5.5 54.8 Oct 2026 Oct 2026 Official leaderboard
31 MiniMax-M3 54.5 Jun 2026 Oct 2026 Official leaderboard
32 Grok 4.7 54.1 Sep 2026 Oct 2026 Official leaderboard
33 MiMo-V2.6 Pro 53.9 Sep 2026 Oct 2026 Official leaderboard
34 GLM-5.3 52.3 Aug 2026 Oct 2026 Official leaderboard
35 Gemini 3.5 Flash-Lite 52.2 Jul 2026 Oct 2026 Official leaderboard
36 DeepSeek V4 Pro 0813 51.6 Aug 2026 Oct 2026 Official leaderboard
37 DeepSeek V4.1 Flash 51.6 Sep 2026 Oct 2026 Official leaderboard
38 aion-labs/aion-3.0-mini 51.2 Oct 2026 Official leaderboard
39 Muse Glimmer 50.0 Aug 2026 Oct 2026 Official leaderboard
40 aion-labs/aion-2.0 49.6 Oct 2026 Official leaderboard
41 Qwen3.8 Flash 46.1 Aug 2026 Oct 2026 Official leaderboard
42 mistralai/mistral-large-4-0 43.2 Oct 2026 Official leaderboard
#ModelScoreSettingReleasedRun onSource
1 GPT-6.2 Astra 73.6 Provisional human rating Nov 2026 Nov 2026 Official leaderboard
2 GPT-6.2 Sol 72.6 Provisional human rating Nov 2026 Nov 2026 Official leaderboard
3 GPT-6.1 Terra 68.3 Oct 2026 Oct 2026 Official leaderboard
4 Claude Opus 5.5 66.3 Sep 2026 Oct 2026 Official leaderboard
5 Gemini 4 Argon 65.1 Sep 2026 Oct 2026 Official leaderboard
6 Gemini 4 Carbon 64.4 Oct 2026 Nov 2026 Official leaderboard
7 Muse Wonder 64.0 Oct 2026 Oct 2026 Official leaderboard
8 Claude Fable 5.1 63.6 Sep 2026 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 63.0 Sep 2026 Oct 2026 Official leaderboard
10 GPT-6 Astra 62.7 Sep 2026 Oct 2026 Official leaderboard
11 Kimi K3.1 Pro 62.5 Oct 2026 Oct 2026 Official leaderboard
12 GPT-6 Sol 62.3 Sep 2026 Oct 2026 Official leaderboard
13 GPT-6 Astra Pro 62.2 Sep 2026 Oct 2026 Official leaderboard
14 Claude Fable 5.2 62.1 Nov 2026 Nov 2026 Official leaderboard
15 GPT-6 Luna 61.3 Sep 2026 Oct 2026 Official leaderboard
16 Muse Spark 1.3 61.0 Sep 2026 Oct 2026 Official leaderboard
17 Claude Opus 4.8 60.7 May 2026 Oct 2026 Official leaderboard
18 Gemini 3.7 Flash 60.3 Aug 2026 Oct 2026 Official leaderboard
19 GPT-6.1 Sol 60.0 Sep 2026 Oct 2026 Official leaderboard
20 Gemini 3.8 Flash 59.7 Sep 2026 Oct 2026 Official leaderboard
21 Kimi K3 58.5 Jul 2026 Oct 2026 Official leaderboard
22 DeepSeek V4.1 Flash 1017 56.8 Oct 2026 Oct 2026 Official leaderboard
23 Qwen 3.8 Max 55.9 prime Aug 2026 Oct 2026 Official leaderboard
24 GLM-5.3-Flash 55.2 Aug 2026 Oct 2026 Official leaderboard
25 MiMo-V2.6 Flash 55.2 Sep 2026 Oct 2026 Official leaderboard
26 Muse Ember 54.9 Nov 2026 Nov 2026 Official leaderboard
27 Claude Haiku 5.5 54.8 Oct 2026 Oct 2026 Official leaderboard
28 MiniMax-M3 54.5 Jun 2026 Oct 2026 Official leaderboard
29 Grok 4.7 54.1 Sep 2026 Oct 2026 Official leaderboard
30 MiMo-V2.6 Pro 53.9 Sep 2026 Oct 2026 Official leaderboard
31 GLM-5.3 52.3 Aug 2026 Oct 2026 Official leaderboard
32 Gemini 3.5 Flash-Lite 52.2 Jul 2026 Oct 2026 Official leaderboard
33 DeepSeek V4 Pro 0813 51.6 Aug 2026 Oct 2026 Official leaderboard
34 DeepSeek V4.1 Flash 51.6 Sep 2026 Oct 2026 Official leaderboard
35 Muse Glimmer 50.0 Aug 2026 Oct 2026 Official leaderboard
36 Qwen3.8 Flash 46.1 Aug 2026 Oct 2026 Official leaderboard

Rankings by part

We read 25 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Story, Intelligence and Prose quality. Claude Fable 5.2 leads 18 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Story

36 models we track have a score here.

  1. 1 Claude Fable 5.2 ↑13 84.1
  2. 2 GPT-6.2 Astra ↓1 81.1
  3. 3 Claude Opus 5.5 ↑1 81.1
  4. 4 GPT-6.2 Sol ↓2 79.1
  5. 5 Claude Fable 5.1 ↑3 79.0
  6. 6 GPT-6 Astra Pro ↑7 78.2
  7. 7 Claude Sonnet 5.5 ↑2 78.2
  8. 8 GPT-6 Astra ↑2 78.1
  9. 9 Muse Wonder ↓2 77.2
  10. 10 Kimi K3.1 Pro ↑1 77.1

The best score, 73.6, costs $8.78 for the full run

Runs on the cost frontier score higher than any run that's cheaper. GPT-6.2 Sol at Provisional human rating comes within 1 of the best for $2.37 for the full run, 4 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Xiaomi Corp OpenAI DeepSeek Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-6.2 Astra, Provisional human rating GPT-6.1 Terra GPT-6 Luna DeepSeek V4.1 Flash 1017 0.00 20.0 40.0 60.0 80.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-6.2 Astra, Provisional human rating GPT-6.1 Terra

Source: Official Fiction Bench (Story Hacker) leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-6.2 Astra Provisional human rating 73.6 $8.78
GPT-6.2 Sol Provisional human rating 72.6 $2.37
GPT-6.1 Terra 68.3 $1.63
Claude Opus 5.5 66.3 $13
Gemini 4 Argon 65.1 $3.51
Gemini 4 Carbon 64.4 $9.19
Muse Wonder 64.0 $13
Claude Fable 5.1 63.6 $32
Claude Sonnet 5.5 63.0 $7.43
GPT-6 Astra 62.7 $14
Kimi K3.1 Pro 62.5 $8.90
GPT-6 Sol 62.3 $3.65
GPT-6 Astra Pro 62.2 $43
Claude Fable 5.2 62.1 $42
GPT-6 Luna 61.3 $0.21
Muse Spark 1.3 61.0 $2.68
Claude Opus 4.8 60.7 $13
Gemini 3.7 Flash 60.3 $1.87
GPT-6.1 Sol 60.0 $2.74
Gemini 3.8 Flash 59.7 $2.88
Kimi K3 58.5 $13
DeepSeek V4.1 Flash 1017 56.8 $0.15
Qwen 3.8 Max prime 55.9 $30
GLM-5.3-Flash 55.2 $0.44
MiMo-V2.6 Flash 55.2 $0.12
Muse Ember 54.9 $0.66
Claude Haiku 5.5 54.8 $0.37
MiniMax-M3 54.5 $0.79
Grok 4.7 54.1 $3.92
MiMo-V2.6 Pro 53.9 $0.40
GLM-5.3 52.3 $8.77
Gemini 3.5 Flash-Lite 52.2 $0.68
DeepSeek V4 Pro 0813 51.6 $1.68
DeepSeek V4.1 Flash 51.6 $0.43
Muse Glimmer 50.0 $0.48
Qwen3.8 Flash 46.1 $0.88

The best score, 73.6, costs $8.78 for the full run

Runs on the cost frontier score higher than any run that's cheaper. GPT-6.2 Sol at Provisional human rating comes within 1 of the best for $2.37 for the full run, 4 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Xiaomi Corp OpenAI DeepSeek Cost frontier One model at each effort level
0.00 20.0 40.0 60.0 80.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-6.2 Astra, Provisional human rating GPT-6.1 Terra GPT-6 Luna DeepSeek V4.1 Flash 1017 0.00 20.0 40.0 60.0 80.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-6.2 Astra, Provisional human rating GPT-6.1 Terra

Source: Official Fiction Bench (Story Hacker) leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-6.2 Astra Provisional human rating 73.6 $8.78
GPT-6.2 Sol Provisional human rating 72.6 $2.37
GPT-6.1 Terra 68.3 $1.63
Claude Opus 5.5 66.3 $13
Gemini 4 Argon 65.1 $3.51
Gemini 4 Carbon 64.4 $9.19
Muse Wonder 64.0 $13
Claude Fable 5.1 63.6 $32
Claude Sonnet 5.5 63.0 $7.43
GPT-6 Astra 62.7 $14
Kimi K3.1 Pro 62.5 $8.90
GPT-6 Sol 62.3 $3.65
GPT-6 Astra Pro 62.2 $43
Claude Fable 5.2 62.1 $42
GPT-6 Luna 61.3 $0.21
Muse Spark 1.3 61.0 $2.68
Claude Opus 4.8 60.7 $13
Gemini 3.7 Flash 60.3 $1.87
GPT-6.1 Sol 60.0 $2.74
Gemini 3.8 Flash 59.7 $2.88
Kimi K3 58.5 $13
DeepSeek V4.1 Flash 1017 56.8 $0.15
Qwen 3.8 Max prime 55.9 $30
GLM-5.3-Flash 55.2 $0.44
MiMo-V2.6 Flash 55.2 $0.12
Muse Ember 54.9 $0.66
Claude Haiku 5.5 54.8 $0.37
MiniMax-M3 54.5 $0.79
Grok 4.7 54.1 $3.92
MiMo-V2.6 Pro 53.9 $0.40
GLM-5.3 52.3 $8.77
Gemini 3.5 Flash-Lite 52.2 $0.68
DeepSeek V4 Pro 0813 51.6 $1.68
DeepSeek V4.1 Flash 51.6 $0.43
Muse Glimmer 50.0 $0.48
Qwen3.8 Flash 46.1 $0.88

3 models have held number one on the official leaderboard since its release in October 2026

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

Anthropic OpenAI
Own leaderboard 3 leaders

Source: Official Fiction Bench (Story Hacker) 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
40.0 50.0 60.0 70.0 80.0 2026 2027 40.0 50.0 60.0 70.0 80.0 2026 2027

Source: Official Fiction Bench (Story Hacker) 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

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (42 entries).
  • Pass Models scored 36 models we track, plus 6 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for GPT-6.2 Astra.
  • Pass Frontier models 10 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on bench.storyhacker.ai: 36 rows for models we track and 6 other entries at the last read.
  • Pass Last read November 20th, 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