MedicalBenchmark
OpenAI: o3 provider

o3

15

#15 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

577.66 pts

Accuracy

97.2%

Correct / Incorrect

583 / 16

Total Cost

$6.40

Overall Performance

(vs. average)
Accuracy

97.2%

avg: 80.6%

Net score

577.66 pts

avg: 453.30 pts

Correct

583

avg: 483

Incorrect

16

avg: 90

Total Cost

$6.40

avg: $9.58

Average response time

8.5s

avg: 17.9s

Output Tokens

481K

avg: 1.3M

Reasoning Tokens

248K

avg: 898K

Average confidence

99.7%

avg: 95.4%

Breakdown by Exam

MIR 2024
27
Correct
193
Incorrect
6
Accuracy
96.5%
Net score
191.00
MIR 2025
26
Correct
191
Incorrect
9
Accuracy
95.5%
Net score
188.00
MIR 2026
5
Correct
199
Incorrect
1
Accuracy
99.5%
Net score
198.66