Minimax Role-Play Bench Full 17-Model Leaderboard 1.0 benchmark leaderboard · SOTA2 Research

Minimax Role-Play Bench Full 17-Model Leaderboard 1.0

84.65Overall ScoreMiniMax-M2-her

Evaluation Results

Method	Links
MiniMax-M2-her2026.01		84.65	—	80.55	79.97	63.99	67.78	89.22	75.3	91.88	91.66	97.51	95.93	97.24	97.15	99.73
GPT-5.12026.01		80.63	—	76.62	72.21	52.18	55.1	81.38	62.49	92.67	89.47	97.05	96.93	96.48	94.9	99.91
Claude-4.5-Opus2026.01		76.62	—	67.23	82.1	57.38	75.33	90.39	78	97.47	94.02	89.9	97.88	99.54	62.23	99.96
Gemini-3-Pro2026.01		75.6	—	62.72	83.87	70.35	74.81	96.33	76.7	94.78	90.25	93.08	99.85	98.5	74.07	99.92
Claude-4.5-Sonnet2026.01		69.35	—	55.72	75.66	52.68	66.23	84.36	72.47	93.94	84.28	90.28	96.55	97.21	67.65	99.71
Gemini-2.5-Pro2026.01		68.23	—	52.36	82.11	66.18	74.14	93.03	78.95	92.78	87.57	86.08	98.53	97.53	48.34	99.92
GPT-4o-2408062026.01		66.39	—	64.96	46.23	27.25	15.41	23.76	37.88	97.01	76.05	89.4	71.18	96.9	89.59	99.94
HER-RL2026.01		65.73	—	59.13	57.74	47.54	41.06	61.99	57.71	69.13	69.03	86.9	74.44	78.42	95.1	99.63
DeepSeek-v3.12026.01		64.22	—	51.11	66.45	49.17	54.94	80.73	56.67	81.83	75.35	88.21	95.78	94.92	62.35	99.8
Claude-3.7-Think2026.01		61.25	—	50.66	59.53	37.97	50.34	65.19	52.07	76.46	75.16	84.15	83.98	83.98	68.64	100
GPT-OSS-120B2026.01		60.72	—	47.27	56.65	31.32	29.06	84.06	41.11	84.66	69.66	91.71	98.16	89.91	79.27	99.49
DeepSeek-v3.22026.01		60.27	—	45.81	66.64	51.22	59.51	76.7	59.13	77.14	76.12	82.83	91.83	94.2	45.33	99.98
HER-SFT2026.01		58.44	—	47.29	52.78	35.95	32.77	55.37	47.55	75.64	69.42	86.4	74.82	77.11	94.1	99.56
GPT-5-Mini2026.01		57.63	—	43.32	50.11	15.43	17.27	55.58	34.53	94.59	83.29	93.78	85.82	93.94	95.45	99.92
Qwen3-32B2026.01		50.76	—	40.38	32.82	24.35	17.66	42.29	33.03	49.77	29.79	89.48	93.07	85.41	80.32	99.11
Grok-4.1-Fast2026.01		48.47	—	29.87	47.51	15.54	22.11	56.86	28.63	89.32	72.59	86.64	98.15	93.03	55.81	99.58
CoSER-70B2026.01		45.38	—	34.32	30.32	25.58	19.46	25.56	41.99	50.35	18.91	82.58	72.11	68.82	90.29	99.1