Conversational Evaluation on CoReflect Iteration 2 (test) benchmark leaderboard · SOTA2 Research

Conversational Evaluation on CoReflect Iteration 2 (test)

4.88ODIGemini 2.5 Pro

Evaluation Results

Method	Links
Gemini 2.5 Pro2026.01		4.88	4.88	4.78	4.85	4.82	4.88	4.84	4.85	4.85
Gemini 2.5 Flash2026.01		4.81	4.77	4.36	4.65	4.72	4.74	4.8	4.75	4.7
Qwen3-Next2026.01		4.77	4.3	4.69	4.59	4.7	4.79	4.78	4.76	4.68
DeepSeek-R12026.01		4.72	4.71	4.63	4.69	4.71	4.66	4.71	4.69	4.69
Claude Sonnet 42026.01		4.7	4.79	4.39	4.63	4.77	4.78	4.79	4.78	4.71
Claude Sonnet 4.52026.01		4.52	4.74	3.99	4.42	4.37	4.73	4.66	4.59	4.51
Claude Haiku 4.52026.01		4.32	4.71	3.75	4.26	4.6	4.66	4.69	4.65	4.46