MedicalBenchmark
OpenAI: GPT-5.2-Codex provider

GPT-5.2-Codex

10

#10 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

579.00 pts

Accuracy

97.3%

Correct / Incorrect

584 / 15

Total Cost

$5.55

Overall Performance

(vs. average)
Accuracy

97.3%

avg: 80.6%

Net score

579.00 pts

avg: 453.30 pts

Correct

584

avg: 483

Incorrect

15

avg: 90

Total Cost

$5.55

avg: $9.58

Average response time

11.4s

avg: 17.9s

Output Tokens

362K

avg: 1.3M

Reasoning Tokens

241K

avg: 898K

Average confidence

99.7%

avg: 95.4%

Breakdown by Exam

MIR 2024
21
Correct
194
Incorrect
6
Accuracy
97.0%
Net score
192.00
MIR 2025
4
Correct
195
Incorrect
4
Accuracy
97.5%
Net score
193.66
MIR 2026
36
Correct
195
Incorrect
5
Accuracy
97.5%
Net score
193.33