MedicalBenchmark
NVIDIA: Llama 3.1 Nemotron Ultra 253B v1 provider

Llama 3.1 Nemotron Ultra 253B v1

173

#173 of 319 models in the general ranking

Cumulative performance across 3 MIR exams

Net score

496.33 pts

Accuracy

86.8%

Correct / Incorrect

521 / 74

Total Cost

$0.77

Overall Performance

(vs. average)
Accuracy

86.8%

avg: 80.6%

Net score

496.33 pts

avg: 453.30 pts

Correct

521

avg: 483

Incorrect

74

avg: 90

Total Cost

$0.77

avg: $9.58

Average response time

12.9s

avg: 17.9s

Output Tokens

330K

avg: 1.3M

Reasoning Tokens

0

avg: 898K

Average confidence

98.8%

avg: 95.4%

Breakdown by Exam

MIR 2024
174
Correct
176
Incorrect
22
Accuracy
88.0%
Net score
168.66
MIR 2025
186
Correct
166
Incorrect
32
Accuracy
83.0%
Net score
155.33
MIR 2026
154
Correct
179
Incorrect
20
Accuracy
89.5%
Net score
172.33