MedicalBenchmark
Meta: Llama 3.3 70B Instruct provider

Llama 3.3 70B Instruct

195

#195 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

478.33 pts

Accuracy

84.0%

Correct / Incorrect

504 / 77

Total Cost

$0.28

Overall Performance

(vs. average)
Accuracy

84.0%

avg: 80.6%

Net score

478.33 pts

avg: 453.30 pts

Correct

504

avg: 483

Incorrect

77

avg: 90

Total Cost

$0.28

avg: $9.58

Average response time

14.1s

avg: 17.9s

Output Tokens

322K

avg: 1.3M

Reasoning Tokens

0

avg: 898K

Average confidence

94.9%

avg: 95.4%

Breakdown by Exam

MIR 2024
201
Correct
171
Incorrect
27
Accuracy
85.5%
Net score
162.00
MIR 2025
203
Correct
161
Incorrect
32
Accuracy
80.5%
Net score
150.33
MIR 2026
177
Correct
172
Incorrect
18
Accuracy
86.0%
Net score
166.00