MedicalBenchmark
Anthropic: Claude Sonnet 4.5 provider

Claude Sonnet 4.5

48

#48 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

562.66 pts

Accuracy

95.3%

Correct / Incorrect

572 / 28

Total Cost

$7.73

Overall Performance

(vs. average)
Accuracy

95.3%

avg: 80.6%

Net score

562.66 pts

avg: 453.30 pts

Correct

572

avg: 483

Incorrect

28

avg: 90

Total Cost

$7.73

avg: $9.58

Average response time

12.9s

avg: 17.9s

Output Tokens

443K

avg: 1.3M

Reasoning Tokens

0

avg: 898K

Average confidence

99.9%

avg: 95.4%

Breakdown by Exam

MIR 2024
22
Correct
194
Incorrect
6
Accuracy
97.0%
Net score
192.00
MIR 2025
66
Correct
185
Incorrect
15
Accuracy
92.5%
Net score
180.00
MIR 2026
57
Correct
193
Incorrect
7
Accuracy
96.5%
Net score
190.66