PlainVerdict
Menu

AI model rankings, by job

Every model we list, ordered by the one public test that best matches each job. Scores are copied from the source, never adjusted.

A ranking is not a recommendation. It orders models by a single test. Our picks also check supporting tests and flag close calls, so the pick for a job is not always first in the table. How the picks are decided.

Write something

Ordered by writing votes. Bar: position among the 77 models listed.

  1. 1Claude Opus 5.5 Our pickAnthropic1,534Ranked 1st of 77
  2. 2Gemini 4 ArgonGoogle DeepMind1,527Ranked 2nd of 77
  3. 3Claude Fable 5.1Anthropic1,512Ranked 3rd of 77
  4. 4Claude Opus 4.6Anthropic1,504Ranked 4th of 77
  5. 5Claude Opus 5Anthropic1,495Ranked 5th of 77
  6. 6Claude Fable 5Anthropic1,493Ranked 6th of 77
  7. 7Gemini 3.8 FlashGoogle DeepMind1,491Ranked 7th of 77
  8. 8Muse WonderMeta1,488Ranked 8th of 77
  9. 9Gemini 3.7 FlashGoogle DeepMind1,487Ranked 9th of 77
  10. 10Claude Opus 4.7Anthropic1,486Ranked 10th of 77

See all 77 ranked for write something

Code something

Ordered by coding votes. Bar: position among the 77 models listed.

  1. 1Claude Opus 5.5 Our pickAnthropic1,547Ranked 1st of 77
  2. 2Gemini 4 ArgonGoogle DeepMind1,538Ranked 2nd of 77
  3. 3MiMo-V2.6-ProXiaomi1,534Ranked 3rd of 77
  4. 4Claude Opus 5Anthropic1,532Ranked 4th of 77
  5. 5Claude Opus 4.6Anthropic1,531Ranked 5th of 77
  6. 6Claude Fable 5.1Anthropic1,530Ranked 6th of 77
  7. 7Claude Opus 4.7Anthropic1,519Ranked 7th of 77
  8. 8Muse WonderMeta1,518Ranked 8th of 77
  9. 9Claude Fable 5Anthropic1,518Ranked 9th of 77
  10. 10Claude Sonnet 5.5Anthropic1,512Ranked 10th of 77

See all 77 ranked for code something

Work through maths

Ordered by hard maths problems. Bar: position among the 43 models listed.

  1. 1GPT-6 AstraOpenAI93.7%Ranked 1st of 43
  2. 1GPT-6.1 Sol Our pickOpenAI93.7%Ranked 1st of 43
  3. 3Claude Opus 5.5Anthropic91.2%Ranked 3rd of 43
  4. 4Claude Fable 5.1Anthropic90.2%Ranked 4th of 43
  5. 5GPT-6 SolOpenAI89.8%Ranked 5th of 43
  6. 6GPT-5.6 SolOpenAI89.1%Ranked 6th of 43
  7. 7Claude Sonnet 5.5Anthropic88.8%Ranked 7th of 43
  8. 8Claude Fable 5Anthropic87.0%Ranked 8th of 43
  9. 9Gemini 4 ArgonGoogle DeepMind86.0%Ranked 9th of 43
  10. 9GPT-5.6 TerraOpenAI86.0%Ranked 9th of 43

See all 43 ranked for work through maths

Get work done

Ordered by business-task votes. Bar: position among the 77 models listed.

  1. 1Gemini 4 Argon Our pickGoogle DeepMind1,515Ranked 1st of 77
  2. 2Claude Opus 5Anthropic1,500Ranked 2nd of 77
  3. 3Claude Opus 5.5Anthropic1,497Ranked 3rd of 77
  4. 4Claude Opus 4.6Anthropic1,497Ranked 4th of 77
  5. 5Claude Fable 5.1Anthropic1,494Ranked 5th of 77
  6. 6Claude Opus 4.7Anthropic1,490Ranked 6th of 77
  7. 7Muse WonderMeta1,487Ranked 7th of 77
  8. 8Muse Spark 1.3Meta1,486Ranked 8th of 77
  9. 9Muse Spark 1.2Meta1,484Ranked 9th of 77
  10. 10Qwen3.8-MaxAlibaba1,480Ranked 10th of 77

See all 77 ranked for get work done

Research a question

Ordered by short factual questions. Bar: position among the 43 models listed.

  1. 1Gemini 4 Argon Our pickGoogle DeepMind76.0%Ranked 1st of 43
  2. 2GPT-6 AstraOpenAI75.6%Ranked 2nd of 43
  3. 3GPT-6.1 SolOpenAI73.9%Ranked 3rd of 43
  4. 4Claude Opus 5.5Anthropic72.2%Ranked 4th of 43
  5. 5Claude Fable 5.1Anthropic70.8%Ranked 5th of 43
  6. 6Claude Fable 5Anthropic70.7%Ranked 6th of 43
  7. 7Gemini 3.8 FlashGoogle DeepMind69.7%Ranked 7th of 43
  8. 7GPT-5.6 SolOpenAI69.7%Ranked 7th of 43
  9. 9Gemini 3.7 FlashGoogle DeepMind69.2%Ranked 9th of 43
  10. 10Gemini 3.5 FlashGoogle DeepMind66.2%Ranked 10th of 43

See all 43 ranked for research a question

Position bars compare only the 77 models listed here. A model with no result for a test is left out of that table and is never shown as low. Data last changed 29 Oct 2026.

Not sure which ranking matters?

Tell us the job. We'll show you the evidence that applies to it.

Find my AI