Benchmark Almanac

DesignArena Spreadsheet

Claude Fable 5.1 1,400 leads since September 2026
Not in model cards

Agents build spreadsheets such as financial models, budgets and trackers as Excel or CSV workbooks. Users compare accuracy, structure and formatting and vote.

One of 40 DesignArena boards. DesignArena ranks entries with Bradley-Terry ratings from user votes.

Kind
Documents and spreadsheets
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
November 2026
Scores refreshed
November 12th, 2026
35
Models scored on its own leaderboard

Claude Fable 5.1 leads with 1,400

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

35 models. Scroll sideways for the rest. Bars start at 900.

  1. 1,400 Claude Fable 5.1
  2. 1,383 GPT-6 Astra
  3. 1,382 Claude Opus 5.5
  4. 1,308 DeepSeek V4.1 Flash
  5. 1,295 Claude Sonnet 5.5
  6. 1,292 Muse Wonder
  7. 1,275 Claude Opus 5
  8. 1,274 Grok 4.6
  9. 1,270 GPT-6.1 Sol
  10. 1,262 Inkling 2 Small
  11. 1,246 MiniMax-M3
  12. 1,223 Muse Spark 1.3
  13. 1,211 Gemini 4 Argon
  14. 1,210 GPT-5.6 Sol
  15. 1,176 Claude Sonnet 5
  16. 1,153 Claude Opus 4.8
  17. 1,153 MiniMax-M3.1-Flash
  18. 1,146 GPT-6.1 Terra
  19. 1,122 Claude Opus 4.6
  20. 1,121 Gemini 3.7 Flash
  21. 1,116 Kimi K3
  22. 1,107 Muse Spark 1.1
  23. 1,094 Gemini 3.8 Flash
  24. 1,093 Gemini 3.6 Flash
  25. 1,093 Muse Spark 1.2
  26. 1,091 Grok 4.5
  27. 1,086 Qwen3.7-Max
  28. 1,064 GLM-5.2
  29. 1,063 Qwen 3.8 Max
  30. 1,062 GPT-5.6 Terra
  31. 1,024 GPT-6 Sol
  32. 1,014 Gemini 3.5 Flash
  33. 1,007 Gemini 3.1 Pro
  34. 964 Grok 4.3
  35. 947 GPT-6 Luna

Bars start at 900.

  1. 1 Claude Fable 5.1 1,400
  2. 2 GPT-6 Astra xhigh 1,383
  3. 3 Claude Opus 5.5 1,382
  4. 4 DeepSeek V4.1 Flash 1017 1,308
  5. 5 Claude Sonnet 5.5 1,295
  6. 6 Muse Wonder max 1,292
  7. 7 Claude Opus 5 1,275
  8. 8 Grok 4.6 1,274
  9. 9 GPT-6.1 Sol 1,270
  10. 10 Inkling 2 Small 1,262
  11. 11 MiniMax-M3 1,246
  12. 12 Muse Spark 1.3 xhigh, max 1,223
  13. 13 Gemini 4 Argon 1,211
  14. 14 GPT-5.6 Sol medium, xhigh 1,210
  15. 15 Claude Sonnet 5 1,176
  16. 16 Claude Opus 4.8 1,153
  17. 17 MiniMax-M3.1-Flash 1,153
  18. 18 GPT-6.1 Terra 1,146
  19. 19 Claude Opus 4.6 1,122
  20. 20 Gemini 3.7 Flash 1,121
  21. 21 Kimi K3 1,116
  22. 22 Muse Spark 1.1 1,107
  23. 23 Gemini 3.8 Flash 1,094
  24. 24 Gemini 3.6 Flash 1,093
  25. 25 Muse Spark 1.2 1,093
  26. 26 Grok 4.5 1,091
  27. 27 Qwen3.7-Max 1,086
  28. 28 GLM-5.2 1,064
  29. 29 Qwen 3.8 Max 1,063
  30. 30 GPT-5.6 Terra 1,062
  31. 31 GPT-6 Sol 1,024
  32. 32 Gemini 3.5 Flash 1,014
  33. 33 Gemini 3.1 Pro 1,007
  34. 34 Grok 4.3 964
  35. 35 GPT-6 Luna 947

Source: Official DesignArena Spreadsheet leaderboard

#ModelScoreSettingDateSource
1 Claude Fable 5.1 1,400 Sep 2026 Official leaderboard
2 GPT-6 Astra 1,383 xhigh Sep 2026 Official leaderboard
3 Claude Opus 5.5 1,382 Sep 2026 Official leaderboard
4 DeepSeek V4.1 Flash 1,308 1017 Sep 2026 Official leaderboard
5 Claude Sonnet 5.5 1,295 Sep 2026 Official leaderboard
6 Muse Wonder 1,292 max Oct 2026 Official leaderboard
7 Claude Opus 5 1,275 Jul 2026 Official leaderboard
8 Grok 4.6 1,274 Aug 2026 Official leaderboard
9 GPT-6.1 Sol 1,270 Sep 2026 Official leaderboard
10 Inkling 2 Small 1,262 Nov 2026 Official leaderboard
11 MiniMax-M3 1,246 Jun 2026 Official leaderboard
12 Muse Spark 1.3 1,223 xhigh, max Sep 2026 Official leaderboard
13 Gemini 4 Argon 1,211 Sep 2026 Official leaderboard
14 GPT-5.6 Sol 1,210 medium, xhigh Jul 2026 Official leaderboard
15 Claude Sonnet 5 1,176 Jun 2026 Official leaderboard
16 Claude Opus 4.8 1,153 May 2026 Official leaderboard
17 MiniMax-M3.1-Flash 1,153 Oct 2026 Official leaderboard
18 GPT-6.1 Terra 1,146 Oct 2026 Official leaderboard
19 Claude Opus 4.6 1,122 Feb 2026 Official leaderboard
20 Gemini 3.7 Flash 1,121 Aug 2026 Official leaderboard
21 Kimi K3 1,116 Jul 2026 Official leaderboard
22 Muse Spark 1.1 1,107 Jul 2026 Official leaderboard
23 Gemini 3.8 Flash 1,094 Sep 2026 Official leaderboard
24 Gemini 3.6 Flash 1,093 Jul 2026 Official leaderboard
25 Muse Spark 1.2 1,093 Aug 2026 Official leaderboard
26 Grok 4.5 1,091 Jul 2026 Official leaderboard
27 Qwen3.7-Max 1,086 May 2026 Official leaderboard
28 GLM-5.2 1,064 Jun 2026 Official leaderboard
29 Qwen 3.8 Max 1,063 Aug 2026 Official leaderboard
30 GPT-5.6 Terra 1,062 Jul 2026 Official leaderboard
31 GPT-6 Sol 1,024 Sep 2026 Official leaderboard
32 Gemini 3.5 Flash 1,014 May 2026 Official leaderboard
33 Gemini 3.1 Pro 1,007 Feb 2026 Official leaderboard
34 Grok 4.3 964 Apr 2026 Official leaderboard
35 GPT-6 Luna 947 Sep 2026 Official leaderboard

5 models have held number one on the official leaderboard since 2023

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

Anthropic MiniMax
Own leaderboard 5 leaders
Claude Fable 5.1 1,400

Source: Official DesignArena Spreadsheet leaderboard

Scores over time

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

On the benchmark’s own leaderboard
900 1,000 1,100 1,200 1,300 1,400 2026 2027 900 1,000 1,100 1,200 1,300 1,400 2026 2027

Source: Official DesignArena Spreadsheet leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Economic work

All 97 in Economic work
  • JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
  • GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate. Seed2.1 Pro 87.9%
  • GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
  • AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Claude Fable 5 75.9%
  • OfficeQA In use Grounded reasoning over a large corpus of U.S. Treasury Bulletin documents: locate tables and compute precise numeric answers. Claude Opus 4.7 86.3%