NVIDIA Nemotron 3 Super 120B A12B BF16 vs Qwen3 235B A22B Instruct 2507
Qwen3 235B A22B Instruct 2507 leads llmrun Score 98.5 to 95.8 (ranges overlap); neither fits 24 GB.
Model A · NVIDIA
NVIDIA Nemotron 3 Super 120B A12B BF16123.6B · 13B active · 256K context · Other · Mar 2026
Model B · Alibaba
Qwen3 235B A22B Instruct 2507235.1B · 22.2B active · 256K context · Apache 2.0 · Jul 2025
Choose NVIDIA Nemotron 3 Super 120B A12B BF16 if you want…
Fits a Mac 64 GB
IQ3_M · 56.1 GB; Qwen3 235B A22B Instruct 2507 needs 72.4 GB at its smallest quant
Lighter on memory
74.7 vs 141.6 GB at Q4_K_M
Newer release
Mar 2026 vs Jul 2025
Choose Qwen3 235B A22B Instruct 2507 if you want…
Higher llmrun Score
98.5 vs 95.8, ranges overlap
Stronger at coding
101.6 vs 95.6
More permissive license
Apache 2.0 vs Other
On your hardware
Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.
8 GB
RTX 4060 · 8 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
12 GB
RTX 3060 12GB · 12 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
16 GB
RTX 5060 Ti 16GB · 16 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
24 GBreference
RTX 4090 · 24 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
32 GB
RTX 5090 · 32 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
48 GB
RTX A6000 · 48 GB tier
NVIDIA Nemotron 3 Super 120B A12B BF16
IQ2_M · ~73 tok/s
42.2 GB · up to 256Kfits
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
Mac 32 GB
Mac Mini M4 · Macs
NVIDIA Nemotron 3 Super 120B A12B BF16
Doesn't fit
smallest quant needs 34.5 GB
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
Mac 64 GB
Mac Studio M4 Max · Macs
NVIDIA Nemotron 3 Super 120B A12B BF16
IQ3_M · ~30 tok/s
56.1 GB · up to 256Kfits
Qwen3 235B A22B Instruct 2507
Doesn't fit
smallest quant needs 72.4 GB
Mac 128 GB
Mac Studio M4 Max · Macs
NVIDIA Nemotron 3 Super 120B A12B BF16
Q4_K_M · ~26 tok/s
74.7 GB · up to 256Kmore headroom
Qwen3 235B A22B Instruct 2507
Q3_K_L · ~21 tok/s
122.4 GB · up to 32K
Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).
Overall
NVIDIA Nemotron 3 Super 120B A12B BF16
Qwen3 235B A22B Instruct 2507
Coding
NVIDIA Nemotron 3 Super 120B A12B BF16
Qwen3 235B A22B Instruct 2507
Science
NVIDIA Nemotron 3 Super 120B A12B BF16
Qwen3 235B A22B Instruct 2507
no data
Reasoning
NVIDIA Nemotron 3 Super 120B A12B BF16
no data
Qwen3 235B A22B Instruct 2507
NVIDIA Nemotron 3 Super 120B A12B BF16 · #60 of 92 open models · #29 of 54 that fit ≤96 GB
Qwen3 235B A22B Instruct 2507 · #51 of 92 open models · #33 of 71 that fit ≤192 GB
Benchmarks
2 shared boards: NVIDIA Nemotron 3 Super 120B A12B BF16 leads 0, Qwen3 235B A22B Instruct 2507 leads 2 · Benchmarks imported Oct 2026
NVIDIA Nemotron 3 Super 120B A12B BF16
1360.3#68/186 open
ArenaLMArena · Oct 2026
Qwen3 235B A22B Instruct 2507
1422.3#32/186 open
ArenaLMArena · Oct 2026
NVIDIA Nemotron 3 Super 120B A12B BF16
38.0%#28/46 open
IndependentEpoch AI · Mar 2026
Qwen3 235B A22B Instruct 2507
38.7%#26/46 open
IndependentEpoch AI · Jul 2025
10 more boards (other shared and single-model boards)
Only Qwen3 235B A22B Instruct 2507 (7)
- ARC-AGI-21.3%
- MMLU-Pro83.0%
- Aider Polyglot59.6%
- ARC-AGI11.0%
- DTBench78.4%
- LMCA23.5%
- Fiction.LiveBench52.9%
VRAM by quantization
Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.
Q2_K
NVIDIA Nemotron 3 Super 120B A12B BF16
53.0 GB54.8 GB file
32K 53.2 GB · 128K 54.3 GB
Qwen3 235B A22B Instruct 2507
100 GB85.7 GB file
32K 103 GB · 128K 113 GB
Q3_K_S
NVIDIA Nemotron 3 Super 120B A12B BF16
54.6 GB61.7 GB file
32K 54.7 GB · 128K 55.9 GB
Qwen3 235B A22B Instruct 2507
103 GB101.4 GB file
32K 106 GB · 128K 116 GB
Q3_K_M
NVIDIA Nemotron 3 Super 120B A12B BF16
60.7 GB61.7 GB file
32K 60.9 GB · 128K 62.0 GB
Qwen3 235B A22B Instruct 2507
115 GB112.4 GB file
32K 118 GB · 128K 128 GB
Q4_0
NVIDIA Nemotron 3 Super 120B A12B BF16
62.3 GB71.0 GB file
32K 62.5 GB · 128K 63.6 GB
Qwen3 235B A22B Instruct 2507
118 GB133.1 GB file
32K 121 GB · 128K 130 GB
Q4_K_M
NVIDIA Nemotron 3 Super 120B A12B BF16
74.7 GB82.5 GB file
32K 74.8 GB · 128K 75.9 GB
Qwen3 235B A22B Instruct 2507
142 GB142.2 GB file
32K 145 GB · 128K 154 GB
Q5_K_M
NVIDIA Nemotron 3 Super 120B A12B BF16
88.6 GB107.3 GB file
32K 88.7 GB · 128K 89.9 GB
Qwen3 235B A22B Instruct 2507
168 GB166.8 GB file
32K 171 GB · 128K 180 GB
Q6_K
NVIDIA Nemotron 3 Super 120B A12B BF16
102 GB114.8 GB file
32K 103 GB · 128K 104 GB
Qwen3 235B A22B Instruct 2507
194 GB193.0 GB file
32K 197 GB · 128K 207 GB
Q8_0
NVIDIA Nemotron 3 Super 120B A12B BF16
124 GB128.5 GB file
32K 124 GB · 128K 125 GB
Qwen3 235B A22B Instruct 2507
236 GB249.9 GB file
32K 239 GB · 128K 248 GB
17 more quants
IQ2_XXS
NVIDIA Nemotron 3 Super 120B A12B BF16
34.5 GB~34.0 GB file
32K 34.7 GB · 128K 35.8 GB
Qwen3 235B A22B Instruct 2507
not listed
IQ2_XS
NVIDIA Nemotron 3 Super 120B A12B BF16
37.6 GB52.0 GB file
32K 37.8 GB · 128K 38.9 GB
Qwen3 235B A22B Instruct 2507
71.0 GB65.6 GB file
32K 74.0 GB · 128K 83.5 GB
IQ2_S
NVIDIA Nemotron 3 Super 120B A12B BF16
39.1 GB52.1 GB file
32K 39.3 GB · 128K 40.4 GB
Qwen3 235B A22B Instruct 2507
74.0 GB66.1 GB file
32K 76.9 GB · 128K 86.4 GB
IQ2_M
NVIDIA Nemotron 3 Super 120B A12B BF16
42.2 GB52.7 GB file
32K 42.4 GB · 128K 43.5 GB
Qwen3 235B A22B Instruct 2507
79.8 GB78.6 GB file
32K 82.8 GB · 128K 92.3 GB
IQ3_XXS
NVIDIA Nemotron 3 Super 120B A12B BF16
48.4 GB56.6 GB file
32K 48.6 GB · 128K 49.7 GB
Qwen3 235B A22B Instruct 2507
91.6 GB93.1 GB file
32K 94.6 GB · 128K 104 GB
IQ3_XS
NVIDIA Nemotron 3 Super 120B A12B BF16
51.5 GB64.5 GB file
32K 51.7 GB · 128K 52.8 GB
Qwen3 235B A22B Instruct 2507
97.5 GB96.9 GB file
32K 100 GB · 128K 110 GB
IQ3_S
NVIDIA Nemotron 3 Super 120B A12B BF16
53.0 GB56.6 GB file
32K 53.2 GB · 128K 54.3 GB
Qwen3 235B A22B Instruct 2507
not listed
IQ3_M
NVIDIA Nemotron 3 Super 120B A12B BF16
56.1 GB66.3 GB file
32K 56.3 GB · 128K 57.4 GB
Qwen3 235B A22B Instruct 2507
106 GB107.3 GB file
32K 109 GB · 128K 119 GB
Q3_K_L
NVIDIA Nemotron 3 Super 120B A12B BF16
63.8 GB66.4 GB file
32K 64.0 GB · 128K 65.1 GB
Qwen3 235B A22B Instruct 2507
121 GB111.2 GB file
32K 124 GB · 128K 133 GB
IQ4_XS
NVIDIA Nemotron 3 Super 120B A12B BF16
66.9 GB64.5 GB file
32K 67.1 GB · 128K 68.2 GB
Qwen3 235B A22B Instruct 2507
127 GB125.5 GB file
32K 130 GB · 128K 139 GB
IQ4_NL
NVIDIA Nemotron 3 Super 120B A12B BF16
70.0 GB64.5 GB file
32K 70.2 GB · 128K 71.3 GB
Qwen3 235B A22B Instruct 2507
133 GB133.1 GB file
32K 136 GB · 128K 145 GB
Q4_1
NVIDIA Nemotron 3 Super 120B A12B BF16
70.0 GB76.6 GB file
32K 70.2 GB · 128K 71.3 GB
Qwen3 235B A22B Instruct 2507
133 GB147.2 GB file
32K 136 GB · 128K 145 GB
Q4_K_S
NVIDIA Nemotron 3 Super 120B A12B BF16
70.0 GB79.0 GB file
32K 70.2 GB · 128K 71.3 GB
Qwen3 235B A22B Instruct 2507
133 GB133.7 GB file
32K 136 GB · 128K 145 GB
Q4_K_L
NVIDIA Nemotron 3 Super 120B A12B BF16
76.2 GB87.4 GB file
32K 76.4 GB · 128K 77.5 GB
Qwen3 235B A22B Instruct 2507
144 GB143.1 GB file
32K 147 GB · 128K 157 GB
Q5_K_S
NVIDIA Nemotron 3 Super 120B A12B BF16
85.5 GB89.7 GB file
32K 85.7 GB · 128K 86.8 GB
Qwen3 235B A22B Instruct 2507
162 GB161.9 GB file
32K 165 GB · 128K 175 GB
Q5_K_L
NVIDIA Nemotron 3 Super 120B A12B BF16
90.1 GB96.9 GB file
32K 90.3 GB · 128K 91.4 GB
Qwen3 235B A22B Instruct 2507
not listed
BF16
NVIDIA Nemotron 3 Super 120B A12B BF16
248 GB241.5 GB file
32K 248 GB · 128K 249 GB
Qwen3 235B A22B Instruct 2507
471 GB470.3 GB file
32K 474 GB · 128K 483 GB
NVIDIA Nemotron 3 Super 120B A12B BF16
Ollama · Q4_K_M
ollama run hf.co/unsloth/NVIDIA-Nemotron-3-Super-120B-A12B-GGUF:Q4_K_MQwen3 235B A22B Instruct 2507
Ollama · Q4_K_M
ollama run hf.co/unsloth/Qwen3-235B-A22B-Instruct-2507-GGUF:Q4_K_M| Spec | NVIDIA Nemotron 3 Super 120B A12B BF16 | Qwen3 235B A22B Instruct 2507 |
|---|---|---|
| Publisher | NVIDIA | Alibaba |
| Parameters | 123.6B | 235.1B |
| Active parameters | 13B per token | 22B per token |
| Architecture | NemotronHForCausalLM | Qwen3MoeForCausalLM |
| Context | 256K tokens | 256K tokens |
| Modalities | Chat, Function-calling | Chat, Function-calling |
| License | Other | Apache 2.0 |
| Released | Mar 2026 | Jul 2025 |
Frequently Asked Questions
Which is better, NVIDIA Nemotron 3 Super 120B A12B BF16 or Qwen3 235B A22B Instruct 2507?
Qwen3 235B A22B Instruct 2507 leads llmrun Score 98.5 to 95.8 (ranges overlap); neither fits 24 GB. Choose NVIDIA Nemotron 3 Super 120B A12B BF16 for fits a Mac 64 GB and lighter on memory; choose Qwen3 235B A22B Instruct 2507 for higher llmrun Score and stronger at coding.
Which needs less VRAM, NVIDIA Nemotron 3 Super 120B A12B BF16 or Qwen3 235B A22B Instruct 2507?
At Q4_K_M, NVIDIA Nemotron 3 Super 120B A12B BF16 requires 74.7 GB while Qwen3 235B A22B Instruct 2507 needs 141.6 GB, making NVIDIA Nemotron 3 Super 120B A12B BF16 the easier model to run on your local setup.
Can I run NVIDIA Nemotron 3 Super 120B A12B BF16 or Qwen3 235B A22B Instruct 2507 on a 16 GB GPU?
NVIDIA Nemotron 3 Super 120B A12B BF16 does not fit (smallest quant needs 34.5 GB); Qwen3 235B A22B Instruct 2507 does not fit (smallest quant needs 72.4 GB).
Which has a longer context window, NVIDIA Nemotron 3 Super 120B A12B BF16 or Qwen3 235B A22B Instruct 2507?
NVIDIA Nemotron 3 Super 120B A12B BF16 supports 262,144 tokens and Qwen3 235B A22B Instruct 2507 supports 262,144 tokens.