Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Open-ended Visual Question Answering on CXR-PA
Loading...
45.32
Accuracy
T2R-R2
19.32
26.07
32.82
39.57
Jun 24, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
T2R-R2
Scale=4B
2026.06
45.32
T2R-R3
Scale=4B
2026.06
43.85
T2R-R1
Scale=4B
2026.06
42.25
Base Model
Scale=4B
2026.06
41.22
Frozen Teacher
Scale=4B
2026.06
38.59
T2R-R3
Scale=2B
2026.06
37.57
T2R-R2
Scale=2B
2026.06
36.98
Judge Reward
Scale=4B
2026.06
36.4
RLVR Only
Scale=4B
2026.06
35.81
T2R-R1
Scale=2B
2026.06
34.64
Frozen Teacher
Scale=2B
2026.06
32.2
Base Model
Scale=2B
2026.06
31.72
Judge Reward
Scale=2B
2026.06
27.04
RLVR Only
Scale=2B
2026.06
20.32
Feedback
Search any
task
Search any
task