Table 1 Characteristics of responses received.

From: Assessing the power of AI: a comparative evaluation of large language models in generating patient education materials in dentistry

Source

Scenario

Flesch reading ease score

Reading level

Average words per sentence

Total sentences

Total words

Chat GPT 4.o

1

52.2

10th to 12th grade (Fairly difficult to read)

10.7

47

505

2

67.6

8th & 9th grade (Plain English)

12.2

43

524

3

60.8

8th & 9th grade (Plain English)

10.5

57

599

4

69.9

8th & 9th grade (Plain English)

9.9

66

655

Claude 3.5 Sonnet

1

57.4

10th to 12th grade (Fairly difficult to read)

5.5

31

171

2

74

7th grade (Fairly easy to read)

5.8

37

213

3

41.7

College (Difficult to read)

4.3

51

218

4

49.8

College (Difficult to read)

4.7

58

274

Gemini 1.5 Flash

1

54.2

10th to 12th grade (Fairly difficult to read)

8.7

63

551

2

54.9

10th to 12th grade (Fairly difficult to read)

8

57

456

3

52.6

10th to 12th grade (Fairly difficult to read)

10.3

51

525

4

52.3

10th to 12th grade (Fairly difficult to read)

10.6

53

563

Llama 3.1–405b

1

53.8

10th to 12th grade (Fairly difficult to read)

9.1

40

362

2

62.1

8th & 9th grade (Plain English)

9.2

39

360

3

76.8

7th grade (Fairly easy to read)

11.4

39

443

4

69.6

8th & 9th grade (Plain English)

9.9

47

467