Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Agentic Task on InstructEdit FineEdit
Loading...
7.56
Speedup
Hybrid Verified Decoding
0.8936
2.6243
4.355
6.0857
May 31, 2026
Speedup
Updated 1d ago
Evaluation Results
Method
Method
Links
Speedup
Hybrid Verified Decoding
Target Model=Llama3.1-...
2026.05
7.56
Hybrid Verified Decoding
Target Model=Llama3.1-...
2026.05
5.49
Hybrid Verified Decoding
Target Model=Qwen3-4B,...
2026.05
4.21
Hybrid Verified Decoding
Target Model=Qwen3-8B,...
2026.05
3.57
Hybrid Verified Decoding
Target Model=Qwen3-4B,...
2026.05
2.67
Hybrid Verified Decoding
Target Model=Qwen3-8B,...
2026.05
2.3
Hybrid Verified Decoding
Target Model=Qwen3-4B,...
2026.05
1.72
Hybrid Verified Decoding
Target Model=Qwen3-8B,...
2026.05
1.51
Hybrid Verified Decoding
Target Model=Qwen3-8B,...
2026.05
1.32
Hybrid Verified Decoding
Target Model=Qwen3-4B,...
2026.05
1.22
Hybrid Verified Decoding
Target Model=Llama3.1-...
2026.05
1.2
Hybrid Verified Decoding
Target Model=Llama3.1-...
2026.05
1.15
Feedback
Search any
task
Search any
task