MedicalBenchmark
OpenAI: GPT-3.5 Turbo Instruct provider

GPT-3.5 Turbo Instruct

281

#281 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

248.33 pts

Accuracy

55.3%

Correct / Incorrect

332 / 251

Total Cost

$0.91

Overall Performance

(vs. average)
Accuracy

55.3%

avg: 80.6%

Net score

248.33 pts

avg: 453.30 pts

Correct

332

avg: 483

Incorrect

251

avg: 90

Total Cost

$0.91

avg: $9.58

Average response time

3.7s

avg: 17.9s

Output Tokens

241K

avg: 1.3M

Reasoning Tokens

0

avg: 898K

Average confidence

96.5%

avg: 95.4%

Breakdown by Exam

MIR 2024
286
Correct
102
Incorrect
93
Accuracy
51.0%
Net score
71.00
MIR 2025
278
Correct
106
Incorrect
88
Accuracy
53.0%
Net score
76.66
MIR 2026
281
Correct
124
Incorrect
70
Accuracy
62.0%
Net score
100.66