Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Science Question Answering on Science QA (Task Acc. & Brier Delta)
Loading...
65
Task Accuracy
FINCH
27.768
37.434
47.1
56.766
May 19, 2026
May 27, 2026
Jun 4, 2026
Jun 13, 2026
Jun 21, 2026
Jun 29, 2026
Jul 8, 2026
Task Accuracy
Brier Delta
Updated 16d ago
Evaluation Results
Method
Method
Links
Task Accuracy
Brier Delta
FINCH
Model=Qwen3-4B
2026.05
65
-
Opus-4.7
2026.07
63
-
DeepSeek-V4-Pro
2026.07
63
-
SFT
Model=Qwen3-4B
2026.05
62.5
-
GPT-5.5
2026.07
60
-
FLOW
Model=Qwen3-4B
2026.05
58.3
-
Kimi-K2.6
2026.07
58
-
L2 Reg
Model=Qwen3-4B
2026.05
57.5
-
WiSE-FT
Model=Qwen3-4B
2026.05
56.7
-
SciReasoner
2026.07
55
-
LoRA
Model=Qwen3-4B
2026.05
51.7
-
STM
Model=Qwen3-4B
2026.05
49.2
-
TALR
Model=Qwen3-4B
2026.05
48.3
-
DFT
Model=Qwen3-4B
2026.05
46.7
-
Base
Model=Qwen3-4B
2026.05
29.2
-
Feedback
Search any
task
Search any
task