排名

模型

平均分

语义理解

闲聊

对话

角色扮演

知识与百科

生成与创作

逻辑与推理

代码

计算

安全

🧝

人类

85.03

90.17

71.53

77.99

82.19

97.44

68.79

90.55

90.45

94.97

86.22

-

gpt-4

70.04

82.91

46.77

66.39

63.46

92.65

66.67

60.33

85.45

61.48

73.02

-

gpt-3.5-turbo

64.4

87.18

45.16

65.57

60.58

85.29

72.36

42.98

72.73

38.52

72.22

🏅️

讯飞星火(v1.5)

63.32

78.26

45.9

59.84

55.88

73.48

54.92

54.7

60

76.86

71.54

-

Claude-2

62.01

83.49

49.59

57.14

52.88

78.68

68.07

53.72

66.06

44.26

65.6

🥈

文心一言(v2.2.0)

61.11

81.9

46.34

56.67

59.8

86.76

47.73

36.52

65.79

52.63

70.63

-

Claude-instant-v1

58.85

76.52

50

58.2

55.77

77.04

61.48

40

66.97

33.61

67.77

🥉

360智脑(4.0)

56.68

76.92

52.46

58.33

54.08

76.8

61.54

37.29

53.64

29.57

67.92

4

ChatGLM2-6B

55.6

74.36

44.35

55.74

56.73

76.47

51.22

40.5

41.82

45.08

66.67

5

internlm-chat-7b

54.85

80.34

48.39

55.74

55.77

77.94

36.59

37.19

51.82

34.43

68.25

6

ChatGLM-130B

53.78

70.94

45.97

56.56

61.54

75.74

55.28

29.75

45.45

31.15

63.49

7

MiniMax-abab5.5

53.61

79.49

45.97

59.84

60.58

85.29

47.97

29.75

30

31.97

61.11

8

通义千问

52.84

74.77

45.97

57.98

53

76.69

38.89

33.06

46.67

39.67

60.4

9

Baichuan-13B-Chat

50.46

64.1

41.94

50

52.88

75

57.72

27.27

40.91

31.15

60.32

10

BELLE-13B

48.71

68.38

46.77

51.64

53.85

64.71

25.2

32.23

48.18

31.97

63.49

11

IDEA-姜子牙-13B-v1.1

47.55

70.09

49.19

48.36

48.08

58.82

32.52

34.71

21.82

45.08

63.49

12

Phoenix-7B

45.39

66.67

41.94

43.44

43.27

55.15

44.72

31.41

36.36

33.61

55.56

13

MOSS-16B

37.01

54.7

39.52

40.16

45.19

35.29

34.96

24.79

32.73

27.05

37.3

14

Llama-2-13B-chat

35.85

52.14

41.94

40.98

32.69

33.82

38.21

28.93

23.64

27.05

38.1

15

Vicuna-13B

34.61

49.57

33.06

32.79

37.5

25.74

30.89

27.27

40.91

35.25

35.71

16

RWKV-7B-World-CHNtuned

30.71

31.62

20.16

22.13

26.92

27.21

23.58

22.31

36.36

60.66

36.51