Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Textual understanding on Arena-Hard
Loading...
67.1
Win Rate
Uni-DPO
0.748
17.974
35.2
52.426
Jun 11, 2025
Win Rate
Updated 8d ago
Evaluation Results
Method
Method
Links
Win Rate
Uni-DPO
Model=Gemma-2-9B-IT
2025.06
67.1
SimPO
Model=Gemma-2-9B-IT
2025.06
59.1
Uni-DPO
Model=Qwen2.5-7B
2025.06
43.5
SFT
Model=Gemma-2-9B-IT
2025.06
40.8
Uni-DPO v0.2
Model=Llama3-8B Instruct
2025.06
40.6
SimPO
Model=Qwen2.5-7B
2025.06
39.5
Uni-DPO
Model=Llama3-8B Instruct
2025.06
37.3
SimPO v0.2
Model=Llama3-8B Instruct
2025.06
36.5
SimPO
Model=Llama3-8B Instruct
2025.06
33.8
DPO
Model=Llama3-8B Instruct
2025.06
32.6
Uni-DPO v0.2
Model=Llama3-8B Base
2025.06
30.7
SimPO v0.2
Model=Llama3-8B Base
2025.06
29.3
SFT
Model=Llama3-8B Instruct
2025.06
25.7
Uni-DPO
Model=Llama3-8B Base
2025.06
23.9
SimPO
Model=Llama3-8B Base
2025.06
23.4
DPO
Model=Llama3-8B Base
2025.06
15.9
SFT
Model=Llama3-8B Base
2025.06
3.3
Feedback
Search any
task
Search any
task