icanwrite.app

AI editing benchmark

Which AI is best for editing?

We hid 90 mistakes in a 6,000-word short story and asked 19 AI models — the GPT models behind ChatGPT, Claude, Gemini and more — to find them

Updated Nov 13, 2026

  • Smartest EditorClaude Fable 5.289/90 caught · $2.29
  • Smart & EfficientGPT-6.1 Sol87/90 caught · $0.20

Mistakes caught by each AI model

Click a model to read every comment it left on the story

  • Grammar & spelling
  • Repetitions
  • Redundancy
  • Consistency
  • Combined: all four checks in one prompt
Issues caughtScoreTimeCost
  1. Claude Fable 5.2Anthropic · Nov 10, 2026
    89/9099%4m 8s$2.29
  2. Claude Fable 5.2 (combined)
    88/9098%3m 45s$1.00
  3. Gemini 4 CarbonGoogle · Nov 10, 2026
    88/9098%3m 53s$0.83
  4. Gemini 4 Carbon (combined)
    84/9093%59s$0.06
  5. Claude Opus 5.5Anthropic · Oct 27, 2026
    88/9098%1m 12s$0.74
  6. Claude Opus 5.5 (combined)
    83/9092%1m 54s$0.33
  7. GPT-6.1 SolOpenAI · Oct 27, 2026
    87/9097%1m 4s$0.20
  8. GPT-6.1 Sol (combined)
    80/9089%1m 9s$0.08
  9. Gemini 4 Hydrogen Google · Nov 13, 2026
    84/9093%42s$0.26
  10. Gemini 4 Hydrogen (combined)
    68/9076%40s$0.11
  11. Gemini 4 ArgonGoogle · Oct 27, 2026
    84/9093%26s$0.12
  12. Gemini 4 Argon (combined)
    80/9089%48s$0.05
  13. GPT-6.1 TerraOpenAI · Oct 27, 2026
    79/9088%29s$0.07
  14. GPT-6.1 Terra (combined)
    76/9084%35s$0.02
  15. GLM 5.3Z.ai · Oct 28, 2026
    78/9087%4m 11s$0.36
  16. GLM 5.3 (combined)
    70/9078%3m 4s$0.12
  17. Gemini 3.8 FlashGoogle · Oct 27, 2026
    77/9086%1m 37s$0.13
  18. Gemini 3.8 Flash (combined)
    72/9080%55s$0.03
  19. Grok 4.6xAI · Oct 28, 2026
    76/9084%4m 40s$0.50
  20. Grok 4.6 (combined)
    66/9073%4m 27s$0.14
  21. Kimi K3Moonshot AI · Oct 28, 2026
    75/9083%58s$0.29
  22. Kimi K3 (combined)
    59/9066%1m 4s$0.07
  23. GPT 5.5OpenAI · Oct 28, 2026
    75/9083%1m 50s$0.94
  24. GPT 5.5 (combined)
    67/9074%1m 50s$0.28
  25. Claude Sonnet 5.5Anthropic · Oct 27, 2026
    73/9081%50s$0.33
  26. Claude Sonnet 5.5 (combined)
    73/9081%1m 32s$0.17
  27. Kimi K 2.6Moonshot AI · Oct 28, 2026
    69/9077%4m 14s$0.43
  28. Kimi K 2.6 (combined)
    49/9054%3m 39s$0.14
  29. Gemini 3.1 ProGoogle · Oct 28, 2026
    67/9074%49s$0.63
  30. Gemini 3.1 Pro (combined)
    51/9057%1m 28s$0.30
  31. Claude Opus 4.8Anthropic · Oct 28, 2026
    65/9072%55s$0.66
  32. Claude Opus 4.8 (combined)
    65/9072%1m 1s$0.19
  33. GPT-6 LunaOpenAI · Oct 27, 2026
    59/9066%1m 20s$0.02
  34. GPT-6 Luna (combined)
    60/9067%1m 49s$0.0049
  35. DeepSeek V4 ProDeepSeek · Oct 28, 2026
    55/9061%1m 20s$0.09
  36. DeepSeek V4 Pro (combined)
    43/9048%21s$0.03
  37. Claude Sonnet 4.6Anthropic · Oct 28, 2026
    55/9061%1m 25s$0.30
  38. Claude Sonnet 4.6 (combined)
    49/9054%1m 44s$0.10

What stood out in the results

The best AI for editing right now is Claude Fable 5.2, which caught 89 of 90 planted mistakes. GPT-6.1 Sol caught 87 at under a tenth of the cost ($0.20 vs $2.29), and Gemini 4 Argon is the best budget pick: 84 caught in 26 seconds for $0.12.

  • Claude Fable 5.2 tops the chart: 89 of 90, every grammar error caught
  • 15 of 18 models do worse on the all-in-one prompt
  • The top five caught all 15 contradictions
  • Above 90%, Gemini 4 Argon is the cheapest, at $0.12
  • Gemini 4 Argon is the fastest top model: 26 seconds
  • GPT-6.1 Terra beats every open-weight model, for $0.07

Free sample

Same checks. Your own draft

The same 4 checks, plus 13 more — from dialogue to point of view. Runs on GPT-6.1 Sol, which caught 87 of 90 mistakes here. See all 17 checks

  • No card needed

Drag & drop your manuscriptUpload your manuscript

.docx, .epub or .md

or
Choose file

Scores for each check

Extra flags are comments that landed on none of the planted mistakes

ModelCaughtGrammar & spellingRepetitionsRedundancyConsistencyAll in oneCommentsExtra flagsTimeCost
Claude Fable 5.289/9045/4514/1515/1515/1588/90−1117174m 8s$2.29
Gemini 4 Carbon88/9043/4515/1515/1515/1584/90−49643m 53s$0.83
Claude Opus 5.588/9044/4514/1515/1515/1583/90−511491m 12s$0.74
GPT-6.1 Sol87/9043/4514/1515/1515/1580/90−79941m 4s$0.20
Gemini 4 Hydrogen 84/9041/4514/1515/1514/1568/90−1689142s$0.26
Gemini 4 Argon84/9042/4513/1514/1515/1580/90−492426s$0.12
GPT-6.1 Terra79/9042/4513/1511/1513/1576/90−389729s$0.07
GLM 5.378/9041/4512/1512/1513/1570/90−8121104m 11s$0.36
Gemini 3.8 Flash77/9041/4510/1512/1514/1572/90−58101m 37s$0.13
Grok 4.676/9042/4512/159/1513/1566/90−1010474m 40s$0.50
Kimi K375/9042/459/1512/1512/1559/90−1691958s$0.29
GPT 5.575/9040/4510/1511/1514/1567/90−89961m 50s$0.94
Claude Sonnet 5.573/9038/4512/1510/1513/1573/90±096350s$0.33
Kimi K 2.669/9039/459/1510/1511/1549/90−2010234m 14s$0.43
Gemini 3.1 Pro67/9040/458/1510/159/1551/90−1687049s$0.63
Claude Opus 4.865/9042/453/1511/159/1565/90±0226655s$0.66
GPT-6 Luna59/9035/4511/155/158/1560/90+16731m 20s$0.02
DeepSeek V4 Pro55/9039/453/154/159/1543/90−1210171m 20s$0.09
Claude Sonnet 4.655/9034/456/156/159/1549/90−618551m 25s$0.30
Download the results (CSV)

Free to reuse with credit to icanwrite, under CC BY 4.0

How we tested the AI editors

Why these four checks

Most editing feedback is taste: two good editors can disagree on pacing or voice, so the only way to score it is another AI's opinion. Grammar, repetitions, redundancy and consistency have right answers — a verb agrees or it doesn't, a name is spelled one way or two — so each mistake can be planted and every model scored the same way.

One story, 90 planted mistakes

The draft is an original 6,000-word short story, clean apart from what we planted in it: 45 grammar and spelling errors, 15 repeated words, 15 redundancies and 15 contradictions. They're the hard kind — agreement across a long subject, a dangling modifier, a real word in the wrong place, a detail that changes chapters later — not typos a spellchecker would catch.

4 checks, one call each

Each model reads the whole story once per check: grammar & spelling, repetitions, redundancy, and consistency. The prompts are adapted from the checks in icanwrite's editor, then frozen, so every model gets exactly the same instructions. Combined asks for all 4 in a single prompt, scored on its own: the gap shows how much a model loses when it's asked to look for everything at once.

How hard each model thinks

Each model gets as much thinking as it can use while still checking a full chapter in a few minutes, since a writer won't wait much longer for comments. For most models that's their default effort, medium or high.

What counts as a catch

A mistake is caught when one of the model's comments lands on it — on either side, for a repetition or a contradiction. Scoring is plain text matching: no AI judges another AI. A comment quoting more than 300 characters catches nothing, so quoting a whole paragraph can't sweep up what's in it.

Time and cost

Time is the slowest of the 4 calls, since they run side by side. Cost is what the calls cost at the provider's API prices. Smart & Efficient is the cheapest model that caught at least 97% of what the smartest did.

Behind the benchmark · 3 min readCan AI be your editor? We tested the best LLMs on 90 mistakesRead the post

Choosing an AI editor

Claude Fable 5.2: it caught all 45 planted errors on the grammar and spelling check, the only model to do so. Claude Opus 5.5 caught 44, Gemini 4 Carbon and GPT-6.1 Sol 43. They were the hard kind: agreement across a long subject, a dangling modifier, a real word in the wrong place. Grammar is also where the models are closest: 15 of the 18 caught at least 39.

See every comment Claude Fable 5.2 left →

Yes. ChatGPT runs on OpenAI's GPT models, and the best one we tested, GPT-6.1 Sol, caught 87 of 90 planted mistakes, two fewer than the top model, for $0.20. GPT-6.1 Terra caught 79 for $0.07. We called the models through OpenAI's API with the same prompts for every model, so the ChatGPT app may answer differently, and which model it uses depends on your plan.

See every comment GPT-6.1 Sol left →

At the top it's close: Anthropic's Claude Fable 5.2 caught 89 of 90, Google's Gemini 4 Carbon 88 and OpenAI's GPT-6.1 Sol 87. The difference is price and speed. Fable cost $2.29 and took four minutes, Carbon $0.83 and just under four, Sol $0.20 and 64 seconds. The cheaper Claude, Opus 5.5, caught 88 for $0.74 in 72 seconds. Claude is the most thorough, GPT the best value, Gemini the fastest at 26 seconds for Gemini 4 Argon's 84.

Gemini 4 Argon: 84 of 90 caught for $0.12 on a 6,000-word story. GPT-6.1 Terra caught 79 for $0.07. Below that the scores fall fast: GPT-6 Luna cost two cents and caught 59.

No. We also gave every model all four checks in a single prompt, and 15 of the 18 caught fewer. Kimi K2.6 lost 20 mistakes, Kimi K3 and Gemini 3.1 Pro 16 each, and even Claude Opus 5.5 lost 5. Ask for one kind of mistake at a time: grammar, then repetitions, then redundancy, then consistency.

The best models can. The consistency check planted 15 contradictions: a ledger that turns from green to red, a grandmother called Ruth and then Rose, a Saturday that's wrong only once you count the days. Claude Fable 5.2, Claude Opus 5.5, Gemini 4 Carbon, GPT-6.1 Sol and Gemini 4 Argon caught all 15. Eight others caught 11 to 14, and five caught 9 or fewer.

Free sample

Ready to check your own draft?

The same 4 checks, plus 13 more — from dialogue to point of view. Runs on GPT-6.1 Sol, which caught 87 of 90 mistakes here. See all 17 checks

  • No card needed

Drag & drop your manuscriptUpload your manuscript

.docx, .epub or .md

or
Choose file

These checks come from icanwrite, a markdown editor for writers

Want to improve the benchmark? Share your ideas