Benchmark Almanac

Corp Speak Arena

GLM-5.3 1,731 leads since August 2026
Not in model cards

Models rewrite text in corporate language. Pairwise votes determine the published Arena rating.

The official API publishes an Arena rating, confidence interval and vote count. Voter identity requires confirmation from the board description.

Made by
Kaggle
Kind
Taste and style
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
September 2026
Scores refreshed
November 12th, 2026
22
Models scored on its own leaderboard

GLM-5.3 leads with 1,731

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.

22 models. Scroll sideways for the rest. Bars start at 1,200.

  1. 1,731 GLM-5.3
  2. 1,696 Claude Opus 5
  3. 1,644 Claude Opus 4.7
  4. 1,616 Claude Opus 5.5
  5. 1,578 Gemini 3.1 Pro
  6. 1,565 Grok 4.7
  7. 1,562 Claude Sonnet 5.5
  8. 1,548 GPT-6.1 Sol
  9. 1,542 Gemini 3.8 Flash
  10. 1,541 Claude Fable 5.1
  11. 1,540 GPT-6 Astra
  12. 1,498 Gemini 4 Argon
  13. 1,481 GPT-5.6 Sol
  14. 1,462 Claude Sonnet 5
  15. 1,460 GPT-5 Pro
  16. 1,457 GPT-5.2
  17. 1,451 GPT-5.6 Terra
  18. 1,441 GPT-5.5
  19. 1,416 Qwen 3.8 Max
  20. 1,367 GPT-5
  21. 1,364 GPT-5.4
  22. 1,218 GPT-6 Luna

Bars start at 1,200.

  1. 1 GLM-5.3 GLM-5.3 Max Reasoning 1,731
  2. 2 Claude Opus 5 Claude Opus 5 Max Reasoning 1,696
  3. 3 Claude Opus 4.7 Claude Opus 4.7 Max Reasoning 1,644
  4. 4 Claude Opus 5.5 Claude Opus 5.5 Max Reasoning 1,616
  5. 5 Gemini 3.1 Pro Gemini 3.1 Pro Preview High Reasoning 1,578
  6. 6 Grok 4.7 Grok 4.7 Xhigh Reasoning 1,565
  7. 7 Claude Sonnet 5.5 Claude Sonnet 5.5 Max Reasoning 1,562
  8. 8 GPT-6.1 Sol GPT-6.1 Sol Max Reasoning 1,548
  9. 9 Gemini 3.8 Flash Gemini 3.8 Flash High Reasoning 1,542
  10. 10 Claude Fable 5.1 Claude Fable 5.1 Max Reasoning 1,541
  11. 11 GPT-6 Astra GPT-6 Astra Max Reasoning 1,540
  12. 12 Gemini 4 Argon Gemini 4 Argon High Reasoning 1,498
  13. 13 GPT-5.6 Sol GPT-5.6 Sol Max Reasoning 1,481
  14. 14 Claude Sonnet 5 Claude Sonnet 5 High Reasoning 1,462
  15. 15 GPT-5 Pro GPT-5 Pro High Reasoning 1,460
  16. 16 GPT-5.2 GPT-5.2 Xhigh Reasoning 1,457
  17. 17 GPT-5.6 Terra GPT-5.6 Terra High Reasoning 1,451
  18. 18 GPT-5.5 GPT-5.5 High Reasoning 1,441
  19. 19 Qwen 3.8 Max Qwen 3.8 Max Xhigh Reasoning 1,416
  20. 20 GPT-5 GPT-5 High Reasoning 1,367
  21. 21 GPT-5.4 GPT-5.4 Xhigh Reasoning 1,364
  22. 22 GPT-6 Luna GPT-6 Luna Max Reasoning 1,218

Source: Official Corp Speak Arena leaderboard

#ModelScoreSettingDateSource
1 GLM-5.3 1,731 GLM-5.3 Max Reasoning Aug 2026 Official leaderboard
2 Claude Opus 5 1,696 Claude Opus 5 Max Reasoning Jul 2026 Official leaderboard
3 Claude Opus 4.7 1,644 Claude Opus 4.7 Max Reasoning Apr 2026 Official leaderboard
4 Claude Opus 5.5 1,616 Claude Opus 5.5 Max Reasoning Sep 2026 Official leaderboard
5 Gemini 3.1 Pro 1,578 Gemini 3.1 Pro Preview High Reasoning Feb 2026 Official leaderboard
6 Grok 4.7 1,565 Grok 4.7 Xhigh Reasoning Sep 2026 Official leaderboard
7 Claude Sonnet 5.5 1,562 Claude Sonnet 5.5 Max Reasoning Sep 2026 Official leaderboard
8 GPT-6.1 Sol 1,548 GPT-6.1 Sol Max Reasoning Sep 2026 Official leaderboard
9 Gemini 3.8 Flash 1,542 Gemini 3.8 Flash High Reasoning Sep 2026 Official leaderboard
10 Claude Fable 5.1 1,541 Claude Fable 5.1 Max Reasoning Sep 2026 Official leaderboard
11 GPT-6 Astra 1,540 GPT-6 Astra Max Reasoning Sep 2026 Official leaderboard
12 Gemini 4 Argon 1,498 Gemini 4 Argon High Reasoning Sep 2026 Official leaderboard
13 GPT-5.6 Sol 1,481 GPT-5.6 Sol Max Reasoning Jul 2026 Official leaderboard
14 Claude Sonnet 5 1,462 Claude Sonnet 5 High Reasoning Jun 2026 Official leaderboard
15 GPT-5 Pro 1,460 GPT-5 Pro High Reasoning Oct 2025 Official leaderboard
16 GPT-5.2 1,457 GPT-5.2 Xhigh Reasoning Dec 2025 Official leaderboard
17 GPT-5.6 Terra 1,451 GPT-5.6 Terra High Reasoning Jul 2026 Official leaderboard
18 GPT-5.5 1,441 GPT-5.5 High Reasoning Apr 2026 Official leaderboard
19 Qwen 3.8 Max 1,416 Qwen 3.8 Max Xhigh Reasoning Aug 2026 Official leaderboard
20 GPT-5 1,367 GPT-5 High Reasoning Aug 2025 Official leaderboard
21 GPT-5.4 1,364 GPT-5.4 Xhigh Reasoning Mar 2026 Official leaderboard
22 GPT-6 Luna 1,218 GPT-6 Luna Max Reasoning Sep 2026 Official leaderboard

6 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.

OpenAI Google DeepMind Anthropic Z.ai
Own leaderboard 6 leaders
GLM-5.3 1,731

Source: Official Corp Speak Arena 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
1,200 1,400 1,600 1,800 2026 1,200 1,400 1,600 1,800 2026

Source: Official Corp Speak Arena 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 Writing and taste

All 30 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