Elo Performance
Overall Elo Ratings
Average of 2000 bootstraps. 95% CI are shown.
Model Size: ≤200B
Open instruct models only
Language Elo Ratings by Model
Average of 2000 bootstraps. 95% CI are shown.
Model Size: ≤200B
Open instruct models only
Model | SEA | MY | TL | ID | MS | TA | TH | VI |
|---|---|---|---|---|---|---|---|---|
 Gemma 4 31B | 1187.1 [-8.35, +8.49] | 1263.8 [-22.17, +23.21] | 1197.9 [-22.10, +22.21] | 1175.8 [-19.92, +19.51] | 1175.8 [-21.78, +21.89] | 1207.1 [-21.70, +21.85] | 1145.2 [-18.01, +18.43] | 1181.0 [-20.87, +19.41] |
 SEA-LION v4.5 (Qwen) 27B AISG | 1128.9 [-6.68, +6.56] | 1175.3 [-17.71, +17.31] | 1108.8 [-16.15, +16.58] | 1105.3 [-14.50, +15.76] | 1104.2 [-18.07, +18.57] | 1184.4 [-18.14, +19.40] | 1130.3 [-15.87, +15.84] | 1126.0 [-16.87, +17.41] |
 Qwen 3.5 122B MoE Alibaba | 1106.3 [-6.17, +6.41] | 1146.5 [-16.10, +15.64] | 1087.6 [-15.13, +15.85] | 1081.0 [-15.12, +15.99] | 1111.9 [-18.14, +19.06] | 1129.4 [-16.53, +17.39] | 1083.4 [-14.66, +14.06] | 1130.2 [-17.11, +18.47] |
Qwen 3.6 27B Alibaba | 1049.4 [-5.58, +5.39] | 1055.7 [-12.77, +14.03] | 1013.6 [-14.41, +14.33] | 1046.7 [-13.01, +13.63] | 1065.4 [-15.22, +15.82] | 1055.2 [-13.40, +13.77] | 1064.4 [-13.70, +13.24] | 1059.0 [-14.91, +14.15] |
 Llama 4 Scout 109B MoE Meta | 1035.2 [-6.93, +6.68] | 1112.5 [-18.49, +19.14] | 1027.6 [-17.85, +17.68] | 1009.8 [-16.56, +16.49] | 1029.1 [-19.56, +20.29] | 1074.5 [-18.58, +18.33] | 1029.3 [-16.06, +15.87] | 990.8 [-17.74, +17.22] |
 Mistral Medium 3.5 128B Mistral AI | 983.5 [-5.60, +5.49] | 907.4 [-11.33, +11.87] | 1017.9 [-14.86, +14.98] | 981.0 [-13.32, +13.81] | 987.6 [-15.46, +16.01] | 1014.3 [-13.02, +13.42] | 973.6 [-12.60, +11.89] | 1001.6 [-13.93, +14.11] |
SEA-LION v4.5 (Gemma E2B) 5B MoE AISG | 911.5 [-6.17, +6.29] | 921.3 [-16.19, +15.31] | 916.8 [-13.75, +14.92] | 909.3 [-14.90, +15.15] | 898.8 [-17.93, +18.44] | 955.0 [-15.50, +16.78] | 883.8 [-14.71, +14.02] | 900.8 [-15.17, +15.53] |
Gemma 4 (E2B) 5B | 897.0 [-6.00, +6.15] | 919.7 [-16.59, +17.76] | 906.3 [-14.74, +15.36] | 876.7 [-15.23, +14.75] | 891.0 [-17.62, +17.71] | 950.2 [-15.47, +16.83] | 863.0 [-15.25, +15.02] | 875.1 [-15.48, +15.94] |
 Nemotron 3 Super 120B MoE NVIDIA | 880.4 [-5.26, +5.33] | 720.8 [-15.09, +15.64] | 894.7 [-14.02, +13.12] | 911.8 [-13.29, +12.65] | 909.2 [-16.04, +15.64] | 796.5 [-15.00, +14.62] | 957.6 [-12.04, +12.28] | 912.5 [-13.61, +14.18] |
 MERaLiON 2 10B A*STAR | 846.8 [-5.50, +5.32] | 741.1 [-15.46, +16.43] | 897.6 [-12.94, +12.86] | 887.1 [-12.76, +12.23] | 847.1 [-15.32, +15.13] | 794.6 [-13.87, +14.20] | 845.5 [-12.64, +13.15] | 876.0 [-12.18, +12.68] |
 Olmo 3.1 32B AI2 | 844.1 [-6.21, +5.76] | 724.9 [-18.50, +18.01] | 878.4 [-15.73, +15.03] | 879.0 [-13.82, +13.72] | 888.9 [-15.31, +15.14] | 765.7 [-16.46, +15.95] | 835.7 [-14.99, +14.64] | 888.7 [-15.19, +16.48] |
 GLM 4.7 Flash 30B MoE Z.ai | 783.7 [-5.45, +5.29] | 790.7 [-15.11, +15.28] | 808.5 [-11.98, +11.80] | 800.7 [-12.35, +12.05] | 796.9 [-12.67, +13.81] | 698.3 [-13.29, +13.34] | 771.0 [-11.99, +11.93] | 787.4 [-12.37, +12.42] |
 Apertus 8B Swiss AI | 743.2 [-5.55, +5.50] | 792.2 [-14.24, +13.80] | 692.6 [-13.30, +13.01] | 721.5 [-12.30, +11.68] | 739.4 [-13.37, +13.33] | 716.7 [-13.35, +13.74] | 781.2 [-12.13, +12.13] | 726.9 [-14.29, +14.33] |
 Tiny Aya Water 3B CohereLabs | 617.8 [-7.21, +7.10] | 504.3 [-27.14, +27.98] | 672.8 [-14.38, +14.04] | 666.8 [-14.17, +14.15] | 687.4 [-16.29, +16.11] | 415.9 [-22.21, +22.09] | 570.6 [-17.61, +17.06] | 657.3 [-15.13, +15.80] |