MedicalBenchmark
NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 provider

Llama 3.3 Nemotron Super 49B V1.5

102

#102 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

539.33 pts

Accuracy

92.2%

Correct / Incorrect

553 / 41

Total Cost

$0.40

Overall Performance

(vs. average)
Accuracy

92.2%

avg: 80.6%

Net score

539.33 pts

avg: 453.30 pts

Correct

553

avg: 483

Incorrect

41

avg: 90

Total Cost

$0.40

avg: $9.58

Average response time

17.8s

avg: 17.9s

Output Tokens

916K

avg: 1.3M

Reasoning Tokens

731K

avg: 898K

Average confidence

98.7%

avg: 95.4%

Breakdown by Exam

MIR 2024
132
Correct
183
Incorrect
14
Accuracy
91.5%
Net score
178.33
MIR 2025
88
Correct
181
Incorrect
18
Accuracy
90.5%
Net score
175.00
MIR 2026
83
Correct
189
Incorrect
9
Accuracy
94.5%
Net score
186.00