Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Graduate-level Science QA on GPQA (Accuracy)
Loading...
52.02
Accuracy
Parallel-Synthesis
2.7032
15.5066
28.31
41.1134
May 14, 2026
May 18, 2026
May 23, 2026
May 28, 2026
Jun 2, 2026
Jun 7, 2026
Jun 12, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Parallel-Synthesis
Backbone=Qwen3-14B, Wo...
2026.06
52.02
Text-Serialization
Backbone=Qwen3-14B, Wo...
2026.06
50
Voting
Backbone=Qwen3-14B, Wo...
2026.06
48.99
KVLINK
Backbone=Qwen3-14B, Wo...
2026.06
35.35
Single
Backbone=Qwen3-14B, Wo...
2026.06
31.31
Qwen2.5-3B-Instruct
Backbone=Qwen2.5-3B, T...
2026.05
30.8
Qwen2.5-3B-RLVR
Backbone=Qwen2.5-3B, T...
2026.05
29.3
Qwen2.5-3B-GRLO
Backbone=Qwen2.5-3B, T...
2026.05
29.3
Qwen2.5-3B-GRLO+RLVR
Backbone=Qwen2.5-3B, T...
2026.05
27.8
Qwen2.5-3B-Base
Backbone=Qwen2.5-3B, T...
2026.05
22.2
CacheBlend
Backbone=Qwen3-14B, Wo...
2026.06
12.63
APE
Backbone=Qwen3-14B, Wo...
2026.06
10.1
Qwen2.5-3B-MathSFT
Backbone=Qwen2.5-3B, T...
2026.05
4.6
Feedback
Search any
task
Search any
task