Benchmark Almanac

Tuesday Work Index

Claude Opus 5.5 76.3 leads since September 2026
Not in model cards

A composite score built from the Surge AI benchmark suite, such as DAYJOB, GDP.pdf, Riemann-bench and ComplexConstraints. It asks whether models get through an ordinary workday.

About 50 models on the board by 2026-10.

Made by
Surge AI
Released
August 2026
Kind
Composite work indices
Level
Expert
Scored
Graded by a model
Answers
Agent task
Tasks
Kept private
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
November 2026
Scores refreshed
November 12th, 2026
44
Models scored on its own leaderboard

Claude Opus 5.5 leads with 76.3

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.

44 models. Scroll sideways for the rest.

  1. 76.3 Claude Opus 5.5
  2. 75.8 Claude Sonnet 5.5
  3. 72.4 Claude Fable 5.1
  4. 72.3 GPT-6 Astra
  5. 70.6 GPT-6.1 Sol
  6. 69.7 Muse Wonder
  7. 67.2 Gemini 4 Argon
  8. 65.0 Claude Opus 5
  9. 64.7 Grok 4.7
  10. 63.9 Muse Spark 1.3
  11. 61.5 Claude Fable 5
  12. 61.1 GPT-5.6 Sol
  13. 60.8 Grok 4.6
  14. 60.2 Kimi K3.1 Pro
  15. 58.9 GPT-6 Sol
  16. 54.9 Qwen 3.8 Max
  17. 54.6 Inkling 2 Small
  18. 54.0 GPT-5.5
  19. 53.8 Tencent Hy4 preview
  20. 53.3 Muse Spark 1.2
  21. 52.7 DeepSeek V4.1 Flash 1017
  22. 50.8 GLM-5.3
  23. 47.9 Gemini 3.7 Flash
  24. 47.3 Kimi K3
  25. 46.1 Claude Sonnet 5
  26. 45.3 GPT-6 Luna
  27. 44.5 Claude Opus 4.8
  28. 43.9 Qwen3.8 Flash
  29. 43.4 Mistral Large 4
  30. 43.3 GPT-5.6 Luna
  31. 43.2 Gemini 3.8 Flash
  32. 42.6 GLM-5.3-Flash
  33. 40.3 Beam 1.0
  34. 39.6 DeepSeek-V4-Pro
  35. 38.0 Claude Opus 4.7
  36. 35.2 DeepSeek V4.1 Flash
  37. 34.3 DeepSeek-V4-Flash-Vision-Exp
  38. 32.2 Inkling
  39. 30.2 Tencent Hy3
  40. 29.5 Gemini 3.1 Pro
  41. 27.6 Muse Glimmer
  42. 25.3 Nemotron 3 Ultra
  43. 24.9 Kimi K2.6
  44. 13.5 Mistral Large 3
  1. 1 Claude Opus 5.5 adaptive/max 76.3
  2. 2 Claude Sonnet 5.5 adaptive/max 75.8
  3. 3 Claude Fable 5.1 adaptive/max 72.4
  4. 4 GPT-6 Astra max 72.3
  5. 5 GPT-6.1 Sol max 70.6
  6. 6 Muse Wonder max 69.7
  7. 7 Gemini 4 Argon high 67.2
  8. 8 Claude Opus 5 adaptive/max 65.0
  9. 9 Grok 4.7 xhigh 64.7
  10. 10 Muse Spark 1.3 max 63.9
  11. 11 Claude Fable 5 adaptive/max 61.5
  12. 12 GPT-5.6 Sol max 61.1
  13. 13 Grok 4.6 xhigh 60.8
  14. 14 Kimi K3.1 Pro max 60.2
  15. 15 GPT-6 Sol max 58.9
  16. 16 Qwen 3.8 Max xhigh 54.9
  17. 17 Inkling 2 Small xhigh 54.6
  18. 18 GPT-5.5 xhigh 54.0
  19. 19 Tencent Hy4 preview high 53.8
  20. 20 Muse Spark 1.2 xhigh 53.3
  21. 21 DeepSeek V4.1 Flash 1017 max 52.7
  22. 22 GLM-5.3 max 50.8
  23. 23 Gemini 3.7 Flash high 47.9
  24. 24 Kimi K3 max 47.3
  25. 25 Claude Sonnet 5 adaptive/max 46.1
  26. 26 GPT-6 Luna max 45.3
  27. 27 Claude Opus 4.8 adaptive/max 44.5
  28. 28 Qwen3.8 Flash xhigh 43.9
  29. 29 Mistral Large 4 high 43.4
  30. 30 GPT-5.6 Luna max 43.3
  31. 31 Gemini 3.8 Flash high 43.2
  32. 32 GLM-5.3-Flash max 42.6
  33. 33 Beam 1.0 max 40.3
  34. 34 DeepSeek-V4-Pro max 39.6
  35. 35 Claude Opus 4.7 adaptive/max 38.0
  36. 36 DeepSeek V4.1 Flash max 35.2
  37. 37 DeepSeek-V4-Flash-Vision-Exp max 34.3
  38. 38 Inkling max 32.2
  39. 39 Tencent Hy3 high 30.2
  40. 40 Gemini 3.1 Pro high 29.5
  41. 41 Muse Glimmer xhigh 27.6
  42. 42 Nemotron 3 Ultra default 25.3
  43. 43 Kimi K2.6 thinking on 24.9
  44. 44 Mistral Large 3 default 13.5

Source: Official Tuesday Work Index leaderboard

#ModelScoreSettingDateSource
1 Claude Opus 5.5 76.3 adaptive/max Sep 2026 Official leaderboard
2 Claude Sonnet 5.5 75.8 adaptive/max Sep 2026 Official leaderboard
3 Claude Fable 5.1 72.4 adaptive/max Sep 2026 Official leaderboard
4 GPT-6 Astra 72.3 max Sep 2026 Official leaderboard
5 GPT-6.1 Sol 70.6 max Sep 2026 Official leaderboard
6 Muse Wonder 69.7 max Oct 2026 Official leaderboard
7 Gemini 4 Argon 67.2 high Sep 2026 Official leaderboard
8 Claude Opus 5 65.0 adaptive/max Jul 2026 Official leaderboard
9 Grok 4.7 64.7 xhigh Sep 2026 Official leaderboard
10 Muse Spark 1.3 63.9 max Sep 2026 Official leaderboard
11 Claude Fable 5 61.5 adaptive/max Jun 2026 Official leaderboard
12 GPT-5.6 Sol 61.1 max Jul 2026 Official leaderboard
13 Grok 4.6 60.8 xhigh Aug 2026 Official leaderboard
14 Kimi K3.1 Pro 60.2 max Oct 2026 Official leaderboard
15 GPT-6 Sol 58.9 max Sep 2026 Official leaderboard
16 Qwen 3.8 Max 54.9 xhigh Aug 2026 Official leaderboard
17 Inkling 2 Small 54.6 xhigh Nov 2026 Official leaderboard
18 GPT-5.5 54.0 xhigh Apr 2026 Official leaderboard
19 Tencent Hy4 preview 53.8 high Aug 2026 Official leaderboard
20 Muse Spark 1.2 53.3 xhigh Aug 2026 Official leaderboard
21 DeepSeek V4.1 Flash 1017 52.7 max Oct 2026 Official leaderboard
22 GLM-5.3 50.8 max Aug 2026 Official leaderboard
23 Gemini 3.7 Flash 47.9 high Aug 2026 Official leaderboard
24 Kimi K3 47.3 max Jul 2026 Official leaderboard
25 Claude Sonnet 5 46.1 adaptive/max Jun 2026 Official leaderboard
26 GPT-6 Luna 45.3 max Sep 2026 Official leaderboard
27 Claude Opus 4.8 44.5 adaptive/max May 2026 Official leaderboard
28 Qwen3.8 Flash 43.9 xhigh Aug 2026 Official leaderboard
29 Mistral Large 4 43.4 high Oct 2026 Official leaderboard
30 GPT-5.6 Luna 43.3 max Jul 2026 Official leaderboard
31 Gemini 3.8 Flash 43.2 high Sep 2026 Official leaderboard
32 GLM-5.3-Flash 42.6 max Aug 2026 Official leaderboard
33 Beam 1.0 40.3 max Oct 2026 Official leaderboard
34 DeepSeek-V4-Pro 39.6 max Apr 2026 Official leaderboard
35 Claude Opus 4.7 38.0 adaptive/max Apr 2026 Official leaderboard
36 DeepSeek V4.1 Flash 35.2 max Sep 2026 Official leaderboard
37 DeepSeek-V4-Flash-Vision-Exp 34.3 max Aug 2026 Official leaderboard
38 Inkling 32.2 max Jul 2026 Official leaderboard
39 Tencent Hy3 30.2 high Jul 2026 Official leaderboard
40 Gemini 3.1 Pro 29.5 high Feb 2026 Official leaderboard
41 Muse Glimmer 27.6 xhigh Aug 2026 Official leaderboard
42 Nemotron 3 Ultra 25.3 default Jun 2026 Official leaderboard
43 Kimi K2.6 24.9 thinking on Apr 2026 Official leaderboard
44 Mistral Large 3 13.5 default Dec 2025 Official leaderboard

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

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

Anthropic
Own leaderboard 3 leaders
Claude Opus 5.5 76.3

Source: Official Tuesday Work Index 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
0 20 40 60 80 2026 2027 0 20 40 60 80 2026 2027

Source: Official Tuesday Work Index 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%