fedi-index: Language: 日本語

Best models for this language by Arena votes in it. Provisional: one source family.

Live data for 2026-10 (recalculated daily) · Methodology v1 · Methodology · History · JSON

All Open weights Closed

#

Tier

Model

fedi-index

Percentile

Value

Sources

9

A

1,468 ±19

94

+44

A

19

A

1,444 ±17

86

+45

A

21

A

1,441 ±16

84

+33

A

24

A

1,436 ±19

82

+44

A

25

A

1,435 ±23

81

+60

A

26

A

1,432 ±16

80

+36

A

27

B

1,429 ±14

80

+27

A

30

B

1,426 ±19

77

+27

A

31

B

1,424 ±15

76

+36

A

32

B

1,423 ±21

76

+105◆

A

37

B

1,421 ±15

72

+80

A

38

B

1,420 ±16

71

+18

A

48

B

1,411 ±17

63

+71

A

49

B

1,410 ±21

62

+38

A

57

B

1,404 ±23

56

+32

A

59

B

1,402 ±21

54

+26

A

61

B

1,402 ±23

53

+46

A

66

C

1,395 ±15

49

+50

A

72

C

1,388 ±21

44

+19

A

76

C

1,376 ±17

41

+50

A

77

C

1,374 ±15

40

-2

A

83

C

1,367 ±15

35

+25

A

86

C

1,362 ±18

33

-2

A

87

C

1,355 ±20

32

+55◆

A

90

C

1,347 ±14

30

-22

A

92

C

1,342 ±19

28

+75

A

94

C

1,339 ±20

27

-40

A

97

D

1,323 ±18

24

-14

A

98

D

1,322 ±16

24

—

A

99

D

1,314 ±15

23

+32

A

100

D

1,309 ±15

22

-29

A

104

D

1,285 ±15

19

-36

A

108

D

1,248 ±25

16

+9

A

119

D

1,159 ±15

7

-236

A

AArena (human votes) ◆ Pareto frontier: nothing is both better and cheaper

fedi-index is an Elo-equivalent on a fixed scale: comparable between months within one methodology version. Tiers S–D: by percentile in the slice, and no better than the confidence interval allows (cut-offs in the methodology). This slice has one source family, so all its scores are provisional. Methodology