MedicalBenchmark
Anthropic: Claude 3.7 Sonnet (thinking) provider

Claude 3.7 Sonnet (thinking)

50

#50 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

561.66 pts

Accuracy

95.2%

Correct / Incorrect

571 / 28

Total Cost

$22.68

Overall Performance

(vs. average)
Accuracy

95.2%

avg: 80.6%

Net score

561.66 pts

avg: 453.30 pts

Correct

571

avg: 483

Incorrect

28

avg: 90

Total Cost

$22.68

avg: $9.58

Average response time

31.3s

avg: 17.9s

Output Tokens

1.4M

avg: 1.3M

Reasoning Tokens

912K

avg: 898K

Average confidence

99.7%

avg: 95.4%

Breakdown by Exam

MIR 2024
56
Correct
191
Incorrect
8
Accuracy
95.5%
Net score
188.33
MIR 2025
62
Correct
186
Incorrect
14
Accuracy
93.0%
Net score
181.33
MIR 2026
50
Correct
194
Incorrect
6
Accuracy
97.0%
Net score
192.00