List your company

Role-playing on CharacterBench 1.0 (test)

4.444MCQwen2.5-7B-Instruct + Character-R1
2.7543.192753.63154.07025Jan 8, 2026

Evaluation results

MethodLinks
Qwen2.5-7B-Instruct + Character-R1Backbone=Qwen2.5-7B, Method=Character-R12026.014.4442.6884.0174.3814.5443.3173.9563.4633.0724.9254.8613.033.1153.832
Llama-3.2-3B-Instruct + Character-R1Backbone=Llama-3.2-3B, Method=Character-R12026.014.2752.1443.9834.5564.953.354.0383.4633.1344.854.7482.942.973.8
Qwen2.5-7B-RoleMRC2026.014.1752.2313.7754.34.7133.2253.6883.1882.9274.754.73533.1053.678
Llama-3.1-8B-RoleMRC2026.014.1692.1633.7174.2314.5883.2253.6063.1252.8644.84.7483.043.063.641
Crab2026.014.1312.1753.64.254.653.1673.6813.3312.9654.854.6843.053.113.665
Haruhi-Zero-7B2026.014.0882.153.5834.2564.5253.1923.6133.4562.9654.854.7732.993.0653.654
Qwen2.5-72B-InstructType=Large Foundation Model2026.014.062.5584.1023.2224.5313.4393.9173.3982.9624.8974.8153.5123.4183.756
Qwen2.5-7B-Instruct + RARBackbone=Qwen2.5-7B, Framework=RAR2026.014.0592.5853.9344.2724.3143.9774.093.1452.8094.8254.8082.8142.9463.737
CoSER-Llama-3.1-8B2026.014.0562.1133.9584.2134.5063.253.73.1562.8834.954.8743.0452.973.667
Qwen2.5-7B-Instruct + SFTBackbone=Qwen2.5-7B, Adaptation=SFT2026.014.022.1653.8734.1054.243.2073.5453.172.7714.724.78833.0563.589
Humanish-Roleplay2026.014.0132.3544.4814.6943.2673.9753.4443.0534.754.8233.013.1453.77
Qwen2.5-7B-Instruct + NeekoBackbone=Qwen2.5-7B, Framework=Neeko2026.013.9972.1923.7753.2884.2573.0843.2123.2292.6934.554.7032.7992.9423.44
Claude-4-opusType=Large Foundation Model2026.013.9972.4354.3984.3624.5083.7513.8773.5873.1764.9364.7353.1513.3513.866
Gemini-3-proType=Large Foundation Model2026.013.9462.9054.0994.1334.1673.5293.6713.3192.9734.824.6333.6133.4233.787
Llama-3-70B-InstructType=Large Foundation Model2026.013.942.593.953.964.393.353.333.062.894.714.743.43.513.678
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B2026.013.9382.3563.84.2754.6313.2333.8563.2383.0974.9184.8613.123.123.726
Llama-3.2-3B-Instruct + RARBackbone=Llama-3.2-3B, Framework=RAR2026.013.8482.013.9784.3734.5114.1123.9183.242.914.7854.7432.7182.9643.701
Llama-3.2-3B-Instruct + NeekoBackbone=Llama-3.2-3B, Framework=Neeko2026.013.7961.8553.8163.4284.5133.1923.4163.3192.7474.5614.5822.6652.9723.451
GPT-4oType=Large Foundation Model2026.013.7932.6473.9784.0274.4843.4143.7233.0462.9744.7634.7713.2613.513.722
CharacterGLMType=Large Foundation Model2026.013.762.183.973.84.032.893.262.942.644.534.513.163.323.461
Baichuan-NPCType=Large Foundation Model2026.013.6722.1344.1324.2164.2543.3664.0223.0013.1794.834.8972.9753.2973.69
Llama-3.2-3B-Instruct + SFTBackbone=Llama-3.2-3B, Adaptation=SFT2026.013.6062.0633.8924.1694.553.53.8943.2193.0034.8884.8233.353.313.713
Llama-3.2-3B-InstructBackbone=Llama-3.2-3B2026.013.5882.0883.8924.34.6063.34.0253.3063.0534.844.6492.893.323.681
GLM-4Type=Large Foundation Model2026.013.5242.3733.7014.1034.383.4873.7283.132.9874.8264.8763.2083.53.679
GPT-3.5-turboType=Large Foundation Model2026.013.492.4513.6924.1554.3453.563.6353.092.8384.7354.7613.6193.7583.702
CharacterGLM-6B2026.013.2452.13.5433.413.3653.13.072.612.54.484.82.842.73.213
Peach-9B-8k-Roleplay2026.012.9312.3443.7173.3883.83.2923.32.8632.8024.8384.7982.4752.4353.306
Peach-2.0-9B-8k-Roleplay2026.012.8192.43.4254.2634.43.8674.0883.353.0284.4754.473.4853.683.673