Vietnamese Performance

Vietnamese Scores by Model

Average of 2000 bootstraps. 95% CI are shown.

1

Legend:

Reasoning model

No open weights

Model Size: All

Open/closed models

Vietnamese Competency & Task Scores

Average of 2000 bootstraps. 95% CI are shown.

Model
VI
Cultural
Inst. Following
Knowledge
Multi-Turn Chat
NLG
NLR
NLU
Safety
Anthropic logo
Anthropic logo
Opus 4.8

Anthropic

78.70

[-1.47, +1.39]

85.42

[-6.31, +5.66]

91.93

[-4.79, +4.37]

88.12

[-3.79, +3.42]

86.96

[-3.45, +3.41]

56.06

[-0.93, +0.91]

83.78

[-2.00, +2.07]

72.93

[-4.14, +4.35]

64.42

[-5.64, +5.83]

Google logo
Google logo
Gemini 3.1 Pro Preview

Google

78.66

[-1.67, +1.49]

79.57

[-9.24, +7.51]

91.87

[-4.72, +4.44]

90.62

[-3.08, +2.92]

84.42

[-3.59, +3.44]

55.86

[-0.80, +0.78]

82.63

[-1.83, +1.87]

72.35

[-4.05, +3.83]

71.98

[-4.79, +4.26]

Alibaba logo
Alibaba logo
Qwen 3.5
397B
MoE

Alibaba

78.15

[-1.56, +1.50]

80.88

[-9.10, +8.06]

96.24

[-3.26, +2.69]

86.20

[-3.82, +3.43]

86.96

[-3.48, +3.21]

55.26

[-0.84, +0.82]

76.58

[-2.28, +2.19]

71.48

[-4.39, +4.26]

71.60

[-3.97, +3.69]

OpenAI logo
OpenAI logo
GPT 6 Astra

OpenAI

77.86

[-1.67, +1.58]

73.57

[-9.43, +9.63]

95.98

[-3.60, +2.83]

92.94

[-2.85, +2.60]

85.28

[-3.64, +3.50]

55.13

[-0.72, +0.76]

82.70

[-2.02, +1.90]

72.41

[-4.23, +4.11]

64.84

[-5.30, +5.03]

Google logo
Google logo
Gemma 4
31B

Google

77.09

[-1.73, +1.64]

75.32

[-9.47, +9.97]

96.21

[-3.35, +2.72]

84.11

[-4.20, +3.85]

87.12

[-3.31, +3.40]

56.21

[-0.91, +0.92]

77.20

[-2.26, +2.25]

68.11

[-4.62, +4.41]

72.42

[-3.89, +3.68]

Z.ai logo
Z.ai logo
GLM 5.2
754B
MoE
FP8

Z.ai

76.93

[-1.45, +1.35]

79.52

[-7.68, +6.86]

95.29

[-3.38, +2.81]

84.93

[-3.18, +3.15]

86.47

[-3.83, +3.49]

53.65

[-0.83, +0.85]

77.39

[-2.03, +1.97]

70.60

[-4.28, +4.19]

67.57

[-4.11, +4.21]

Alibaba logo
Alibaba logo
Qwen 3.6
27B

Alibaba

76.12

[-1.46, +1.39]

75.69

[-7.74, +6.89]

94.40

[-3.92, +3.34]

83.69

[-3.82, +3.48]

84.71

[-3.30, +3.19]

54.67

[-0.81, +0.78]

79.22

[-2.18, +1.96]

71.71

[-4.22, +3.84]

64.91

[-4.28, +3.85]

AISG logo
AISG logo
SEA-LION v4.5 (Qwen)
27B

AISG

73.85

[-1.40, +1.40]

74.95

[-6.98, +6.52]

94.92

[-2.89, +2.35]

81.53

[-3.95, +3.72]

79.32

[-4.05, +4.01]

51.78

[-0.78, +0.76]

77.49

[-2.13, +2.11]

68.00

[-4.15, +4.07]

62.77

[-4.01, +3.77]

Meta logo
Meta logo
Muse Glimmer
30B

Meta

73.52

[-1.47, +1.38]

61.37

[-7.98, +7.60]

95.88

[-3.26, +2.34]

81.49

[-4.24, +4.01]

87.09

[-2.85, +2.82]

53.87

[-0.76, +0.73]

71.21

[-2.27, +2.33]

69.61

[-4.30, +4.22]

67.62

[-4.02, +3.78]

DeepSeek logo
DeepSeek logo
DeepSeek V4 Pro
1600B
MoE

DeepSeek

73.40

[-1.56, +1.54]

56.03

[-7.92, +8.39]

90.42

[-4.95, +4.46]

87.02

[-3.43, +3.23]

86.77

[-3.18, +3.07]

55.96

[-0.87, +0.84]

72.98

[-2.20, +2.16]

70.40

[-3.99, +3.95]

67.66

[-4.08, +4.04]

Minimax logo
Minimax logo
MiniMax M3
428B
MoE

Minimax

73.40

[-1.72, +1.71]

69.99

[-8.85, +8.62]

85.16

[-5.75, +5.08]

83.55

[-3.59, +3.49]

86.14

[-3.68, +3.55]

54.50

[-0.72, +0.68]

74.85

[-2.06, +2.13]

68.56

[-4.25, +4.03]

64.41

[-4.19, +3.95]

Thinking Machines logo
Thinking Machines logo
Inkling Small
276B
MoE

Thinking Machines

72.52

[-1.40, +1.43]

52.86

[-7.85, +8.49]

95.05

[-3.51, +2.92]

84.88

[-4.17, +3.66]

89.30

[-2.97, +2.77]

53.61

[-0.70, +0.67]

69.65

[-2.11, +2.07]

70.85

[-4.04, +4.00]

63.93

[-3.78, +3.70]

AISG logo
AISG logo
SEA-LION v4.8 (Nemotron)
120B
MoE

AISG

70.36

[-1.65, +1.67]

67.63

[-8.61, +8.08]

82.03

[-6.32, +5.71]

79.64

[-4.14, +4.07]

84.88

[-3.15, +3.10]

53.05

[-0.77, +0.80]

73.88

[-2.32, +2.20]

67.34

[-4.32, +4.08]

54.40

[-4.21, +3.87]

OpenAI logo
OpenAI logo
GPT OSS
120B
MoE
mxfp4

OpenAI

69.81

[-1.69, +1.59]

66.08

[-7.69, +7.34]

88.25

[-5.16, +4.49]

74.79

[-4.46, +4.42]

83.50

[-3.98, +3.64]

53.40

[-0.71, +0.69]

71.74

[-2.38, +2.49]

65.06

[-4.17, +3.94]

55.67

[-4.45, +4.75]

Mistral AI logo
Mistral AI logo
Mistral Medium 3.5
128B

Mistral AI

68.82

[-1.47, +1.43]

46.55

[-6.47, +6.56]

86.44

[-5.25, +4.64]

77.99

[-3.95, +4.10]

82.88

[-3.98, +3.94]

55.73

[-0.90, +0.90]

69.46

[-2.18, +2.08]

71.23

[-3.86, +3.85]

60.30

[-4.30, +3.97]

NVIDIA logo
NVIDIA logo
Nemotron 3 Ultra
550B
MoE

NVIDIA

68.57

[-1.74, +1.79]

66.25

[-8.58, +8.90]

77.42

[-7.07, +6.27]

81.64

[-3.72, +3.74]

88.38

[-3.02, +2.92]

53.54

[-0.73, +0.72]

67.19

[-2.05, +2.10]

68.14

[-4.40, +4.10]

46.01

[-4.94, +5.47]

CohereLabs logo
CohereLabs logo
Command A+
218B
MoE

CohereLabs

66.44

[-1.51, +1.56]

60.80

[-7.39, +8.14]

88.13

[-4.56, +3.89]

74.55

[-4.55, +4.07]

81.72

[-3.67, +3.65]

36.39

[-0.76, +0.73]

70.96

[-2.07, +2.11]

64.33

[-4.24, +3.89]

54.66

[-4.37, +4.24]

AISG logo
AISG logo
SEA-LION v4.8 (Nemotron)
30B
MoE

AISG

64.86

[-1.67, +1.63]

50.84

[-7.97, +8.65]

78.37

[-6.11, +5.79]

74.65

[-4.61, +4.18]

83.06

[-3.60, +3.29]

52.08

[-0.78, +0.84]

66.52

[-2.17, +2.19]

65.85

[-4.20, +3.98]

47.50

[-4.18, +4.07]

Swiss AI logo
Swiss AI logo
Apertus 1.5
70B

Swiss AI

64.66

[-1.62, +1.63]

62.81

[-7.24, +7.51]

83.05

[-5.08, +4.45]

61.50

[-4.70, +4.38]

68.25

[-4.93, +4.72]

53.48

[-0.72, +0.74]

66.15

[-2.40, +2.38]

68.00

[-4.02, +3.97]

54.04

[-4.29, +4.30]

AISG logo
AISG logo
SEA-LION v4.5 (Gemma E2B)
5B
MoE

AISG

61.95

[-1.75, +1.78]

50.31

[-7.80, +8.56]

85.92

[-5.56, +5.04]

55.50

[-5.09, +5.33]

76.67

[-4.80, +4.41]

53.25

[-0.70, +0.65]

61.11

[-2.78, +2.60]

64.01

[-4.31, +4.02]

48.80

[-4.95, +4.88]

Google logo
Google logo
Gemma 4 (E2B)
5B

Google

61.19

[-1.66, +1.69]

45.89

[-6.21, +6.56]

86.79

[-5.60, +5.00]

55.54

[-5.21, +5.50]

77.49

[-4.92, +4.49]

52.37

[-0.64, +0.60]

61.42

[-2.76, +2.56]

62.12

[-4.52, +4.22]

47.86

[-5.05, +5.01]

NVIDIA logo
NVIDIA logo
Nemotron 3.5 Lightning
30B
MoE

NVIDIA

60.97

[-1.55, +1.52]

35.49

[-5.04, +5.37]

75.68

[-6.76, +6.35]

72.44

[-4.40, +4.47]

83.06

[-3.51, +3.59]

48.94

[-0.69, +0.67]

61.21

[-2.51, +2.39]

63.49

[-4.11, +4.00]

47.42

[-4.70, +4.67]

NVIDIA logo
NVIDIA logo
Nemotron 3 Super
120B
MoE

NVIDIA

55.31

[-1.48, +1.55]

62.76

[-7.88, +7.88]

80.78

[-5.18, +4.94]

50.82

[-3.65, +3.68]

83.20

[-3.11, +3.01]

51.89

[-0.70, +0.70]

40.11

[-2.28, +2.32]

36.31

[-3.73, +3.54]

36.61

[-4.22, +3.79]

A*STAR logo
A*STAR logo
MERaLiON 2
10B

A*STAR

51.89

[-1.81, +1.85]

50.65

[-7.78, +8.30]

63.93

[-6.90, +6.43]

27.41

[-5.24, +5.43]

50.51

[-4.86, +4.97]

52.58

[-0.73, +0.71]

56.05

[-3.12, +2.91]

55.45

[-4.04, +3.85]

58.56

[-5.29, +4.88]

AI2 logo
AI2 logo
Olmo 3.1
32B

AI2

51.07

[-1.89, +2.01]

33.95

[-9.34, +9.17]

81.14

[-6.73, +5.89]

31.23

[-5.94, +5.81]

69.95

[-4.87, +4.84]

51.36

[-0.77, +0.76]

41.53

[-3.72, +3.58]

64.92

[-4.02, +4.23]

34.47

[-4.80, +5.26]

No records

Legend:

Reasoning model

No open weights

Model Size: All

Open/closed models