Emotion Analysis on MISID benchmark leaderboard · SOTA2 Research

Emotion Analysis on MISID

90.75Emotion State AccuracyClaude-Sonnet-4.5

Evaluation Results

Method	Links
Claude-Sonnet-4.52026.04		90.75	81.46
Claude-Sonnet-4.52026.04		89.01	81.87
Grok-4-Fast2026.04		88.92	80.62
Qwen3.5-Plus2026.04		87.32	81.08
Qwen3.5-Plus2026.04		87.16	83.55
Gemini-3-flash2026.04		87.09	85.1
Grok-4-Fast2026.04		86.59	81.1
Grok-4-Fast2026.04		86.21	79.89
GPT-5.12026.04		84.14	74.59
GLM-4.72026.04		83.79	81.03
Qwen3.5-Plus2026.04		83.78	80.28
DeepSeek-R12026.04		83.73	80.18
DeepSeek-R12026.04		83.45	78.1
GLM-4.72026.04		83.39	81.3
GPT-5.12026.04		82.73	74.32
Gemini-2.5-flash2026.04		82.55	78.55
MiniMax-M2.12026.04		82.02	82.96
GPT-4o2026.04		80.28	75.7
MiniMax-M2.12026.04		80.08	82.27
Qwen3-Max2026.04		80.03	78.2
GPT-4o2026.04		79.87	75.9
Qwen3-VL-235B-A22B2026.04		79.05	76.19
Qwen3-Max2026.04		77.18	77.52
GPT-4o2026.04		76.43	73.75
Qwen3-VL-Max2026.04		74.05	75.57
Qwen3-VL-Plus2026.04		72.83	61.78
DeepSeek-V32026.04		69.53	71.2
Qwen3-VL-Flash2026.04		69.23	46.7
DeepSeek-V32026.04		66.91	71.4