Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Dialogue Evaluation on Arena-Hard
Loading...
3.29
Accepted Length (τ)
DSpark
2.4476
2.6663
2.885
3.1037
Jul 6, 2026
Accepted Length (τ)
Updated 18d ago
Evaluation Results
Method
Method
Links
Accepted Length (τ)
DSpark
target_model=Qwen3-4B
2026.07
3.29
DSpark
target_model=Qwen3-8B
2026.07
3.21
DSpark
target_model=Qwen3-14B
2026.07
3.13
DSpark
target_model=Gemma4-12B
2026.07
2.92
DFlash
target_model=Qwen3-4B
2026.07
2.83
DFlash
target_model=Qwen3-8B
2026.07
2.81
DFlash
target_model=Qwen3-14B
2026.07
2.72
Eagle3
target_model=Gemma4-12B
2026.07
2.72
DFlash
target_model=Gemma4-12B
2026.07
2.59
Eagle3
target_model=Qwen3-4B
2026.07
2.55
Eagle3
target_model=Qwen3-8B
2026.07
2.54
Eagle3
target_model=Qwen3-14B
2026.07
2.48
Feedback
Search any
task
Search any
task