Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Retrosynthesis on Retrosynthesis
Loading...
100
Validity
Qwen3-4B w SciDC
46.44
60.345
74.25
88.155
Apr 8, 2026
Validity
Hit@1
Updated 9d ago
Evaluation Results
Method
Method
Links
Validity
Hit@1
Qwen3-4B w SciDC
Backbone model=Qwen3-4...
2026.04
100
52.2
Qwen3-14B w SciDC
Backbone model=Qwen3-1...
2026.04
100
41.3
Claude-3.5
Backbone model=Claude-3.5
2026.04
92.8
47.3
Qwen3-4B
Backbone model=Qwen3-4B
2026.04
84.6
25.4
Qwen3-14B
Backbone model=Qwen3-14B
2026.04
78.1
31.8
GPT-5
Backbone model=GPT-5
2026.04
64.7
31.8
Qwen3-4B w/o K
Backbone model=Qwen3-4...
2026.04
63.5
0
Qwen3-14B w/o K
Backbone model=Qwen3-1...
2026.04
48.5
0
Feedback
Search any
task
Search any
task