Compare · ModelsLive · 2 picked · head to head
gpt-oss-20b vs Nemotron 3 Ultra
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Nemotron 3 Ultra wins on 16/16 benchmarks
Nemotron 3 Ultra wins 16 of 16 shared benchmarks. Leads in speed · knowledge · general.
Category leads
speed·Nemotron 3 Ultraknowledge·Nemotron 3 Ultrageneral·Nemotron 3 Ultramath·Nemotron 3 Ultracoding·Nemotron 3 Ultra
Hype vs Reality
Attention vs performance
gpt-oss-20b
#101 by perf·no signal
Nemotron 3 Ultra
#167 by perf·#8 by attention
Best value
gpt-oss-20b
32.8x better value than Nemotron 3 Ultra
gpt-oss-20b
983.3 pts/$
$0.05/M
Nemotron 3 Ultra
30.0 pts/$
$1.50/M
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
NVIDIA
$5.24T·Big Tech
Head to head
16 benchmarks · 2 models
gpt-oss-20bNemotron 3 Ultra
Artificial Analysis · CritPt
Nemotron 3 Ultra leads by +1.7
gpt-oss-20b
1.4
Nemotron 3 Ultra
3.1
Artificial Analysis · GDPval
Nemotron 3 Ultra leads by +25.7
gpt-oss-20b
0.0
Nemotron 3 Ultra
25.7
Artificial Analysis · GPQA Diamond
Nemotron 3 Ultra leads by +17.9
gpt-oss-20b
68.8
Nemotron 3 Ultra
86.7
Artificial Analysis · Humanity's Last Exam
Nemotron 3 Ultra leads by +17.4
gpt-oss-20b
11.0
Nemotron 3 Ultra
28.4
Artificial Analysis · IFBench
Nemotron 3 Ultra leads by +16.3
gpt-oss-20b
65.1
Nemotron 3 Ultra
81.4
Artificial Analysis · Long Context Reasoning
Nemotron 3 Ultra leads by +44.6
gpt-oss-20b
34.7
Nemotron 3 Ultra
79.3
Artificial Analysis · Quality Index
Nemotron 3 Ultra leads by +14.0
gpt-oss-20b
9.0
Nemotron 3 Ultra
22.9
Artificial Analysis · SciCode
Nemotron 3 Ultra leads by +1.4
gpt-oss-20b
38.9
Nemotron 3 Ultra
40.3
Artificial Analysis · tau2-Bench Telecom
Nemotron 3 Ultra leads by +23.1
gpt-oss-20b
60.2
Nemotron 3 Ultra
83.3
Artificial Analysis · Terminal-Bench Hard
Nemotron 3 Ultra leads by +25.8
gpt-oss-20b
10.6
Nemotron 3 Ultra
36.4
Chess Puzzles
Nemotron 3 Ultra leads by +7.4
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
gpt-oss-20b
0.0
Nemotron 3 Ultra
7.4
Dtbench
Nemotron 3 Ultra leads by +36.9
gpt-oss-20b
46.7
Nemotron 3 Ultra
83.5
GPQA diamond
Nemotron 3 Ultra leads by +32.7
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
gpt-oss-20b
47.7
Nemotron 3 Ultra
80.5
Lmca
Nemotron 3 Ultra leads by +26.3
gpt-oss-20b
17.1
Nemotron 3 Ultra
43.4
OTIS Mock AIME 2024-2025
Nemotron 3 Ultra leads by +21.4
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
gpt-oss-20b
65.2
Nemotron 3 Ultra
86.7
WeirdML
Nemotron 3 Ultra leads by +2.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
gpt-oss-20b
40.9
Nemotron 3 Ultra
43.5
Full benchmark table
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
Artificial Analysis · CritPt | 1.4 | 3.1 |
Artificial Analysis · GDPval | 0.0 | 25.7 |
Artificial Analysis · GPQA Diamond | 68.8 | 86.7 |
Artificial Analysis · Humanity's Last Exam | 11.0 | 28.4 |
Artificial Analysis · IFBench | 65.1 | 81.4 |
Artificial Analysis · Long Context Reasoning | 34.7 | 79.3 |
Artificial Analysis · Quality Index | 9.0 | 22.9 |
Artificial Analysis · SciCode | 38.9 | 40.3 |
Artificial Analysis · tau2-Bench Telecom | 60.2 | 83.3 |
Artificial Analysis · Terminal-Bench Hard | 10.6 | 36.4 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 0.0 | 7.4 |
Dtbench | 46.7 | 83.5 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 47.7 | 80.5 |
Lmca | 17.1 | 43.4 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 65.2 | 86.7 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 40.9 | 43.5 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.02 | $0.09 | 131K tokens (~66 books) | $0.36 | |
| $0.60 | $2.40 | 262K tokens (~131 books) | $10.50 |