Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Open-ended generation on Vicuna
Loading...
99.1
Skywork Reward V2 Score
Student
28.484
46.817
65.15
83.483
Apr 21, 2026
Skywork Reward V2 Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Skywork Reward V2 Score
Student
Base Model=Gemma-3-4B
2026.04
99.1
Distillable
Base Model=Gemma-3-12B
2026.04
99
Teacher
Base Model=Gemma-3-12B
2026.04
98.9
Undistillable
Base Model=Gemma-3-12B
2026.04
98.9
Teacher
Base Model=Qwen3-8B
2026.04
96.7
Distillable
Base Model=Qwen3-8B
2026.04
96.6
Undistillable
Base Model=Qwen3-8B
2026.04
96.5
Improve
Base Model=Gemma-3-4B
2026.04
90.2
Student
Base Model=Qwen3-1.7B
2026.04
89.1
Improve
Base Model=Qwen3-1.7B
2026.04
87.4
GKD-FKL
Base Model=Qwen3-1.7B
2026.04
86.3
GKD-FKL
Base Model=Gemma-3-4B
2026.04
85.7
SFT
Base Model=Gemma-3-4B
2026.04
82.1
GKD-RKL
Base Model=Gemma-3-4B
2026.04
81.5
SFT
Base Model=Qwen3-1.7B
2026.04
81.4
GKD-RKL
Base Model=Qwen3-1.7B
2026.04
76.4
Misled
Base Model=Gemma-3-4B
2026.04
32.9
Misled
Base Model=Qwen3-1.7B
2026.04
31.2
Feedback
Search any
task
Search any
task