Multilingual Evaluation on In-House Multilingual Datasets SEA Average 1.0 benchmark leaderboard · SOTA2 Research

Multilingual Evaluation on In-House Multilingual Datasets SEA Average 1.0

90.15en ScoreGPT-5-Thinking

Evaluation Results

Method	Links
GPT-5-Thinking2025.12		90.15	89.58	88.08	87.51	87.67	85.96	83.97	86.64
gemini-2.5-pro2025.12		89.39	88.84	86.62	88.28	89.13	86.55	85.76	87.27
Qwen3-235B-A22B2025.12		88.73	88.44	86.76	87.31	87.46	84.83	83.26	85.92
GLM4.52025.12		88.51	88.11	85.31	86.61	87.82	82.12	84.09	85.19
CompassMax-V3-Thinking2025.12		88.07	85.64	87.89	86.56	86.84	86.04	84.75	86.41
CompassMax-V32025.12		87.96	83.08	88.71	84.63	84.54	85.66	84.38	85.58
GPT-4o2025.12		87.31	88.32	80.02	86.17	85.84	85.07	80.99	83.62
Deepseek R12025.12		85.96	86.38	85.08	83.39	84.02	83.73	84.34	84.11