Nemotron 3 Ultra 550B A55B vs NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
Wins 1 of 4 areas
Following instructions
Wins 3 of 4 areas
Coding · Agents · Reasoning
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 is the stronger all-rounder.Nemotron 3 Ultra 550B A55B is better at following instructions.
Scores updated · 22 tests both models report · How we compare
Where each one wins
Tests won in each of the four areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software02NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF162 of 5 tests · 3 ties
- ReasoningHard problems that need careful thinking02NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF162 of 2 tests
- AgentsCarrying out multi-step tasks on its own01NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF161 of 1 test
- Following instructionsDoing exactly what it is asked10Nemotron 3 Ultra 550B A55B1 of 1 test
Agents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+33.6points ahead
- Research-level physics problemsCritPt+7.8points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+6.8points ahead
Where Nemotron 3 Ultra 550B A55B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+10.5points ahead
Every test, side by side
All 22 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
- Terminal-Bench 2.1NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 6.853.960.7+6.8
- SWE-bench ProNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 2.149.751.8+2.1
- LiveCodeBench v6tie8988.2tie
- SWE-bench Verifiedtie70.771.1tie
- SWE-bench Multilingualtie67.767.7tie
AgentsNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
- BrowseCompNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 33.644.478+33.6
ReasoningNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
- CritPtNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 7.83.110.9+7.8
- GPQA DiamondNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 2.886.789.5+2.8
Following instructionsNemotron 3 Ultra 550B A55B
- IFBenchNemotron 3 Ultra 550B A55B by 10.581.470.9+10.5
Other results13 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- BrowseComp [offline]NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 344478+34
- DeepSWE [offline]NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 21.7324.6+21.7
- LMArena v3 proxyNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 18.434.352.7+18.4
- tau2-bench telecomNemotron 3 Ultra 550B A55B by 9.88373.2+9.8
- GDPval-AA v2NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 63 rating points11481211+63 rating
- MMLU-ProXNemotron 3 Ultra 550B A55B by 6.18376.9+6.1
- SciCode, subtaskNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 5.14247.1+5.1
- tau2-bench retailNemotron 3 Ultra 550B A55B by 4.573.769.2+4.5
- tau2-bench airlineNemotron 3 Ultra 550B A55B by 3.379.876.5+3.3
- HLE, text-onlyNVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 2.629.231.8+2.6
- WMT24++NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 1.183.784.8+1.1
- Terminal Bench 4 [offline]NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 by 101+1
- τ³-Bench Bankingtie18.418.9tie
Questions people ask
Which is better, Nemotron 3 Ultra 550B A55B or NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16?
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 wins three of the four areas where both have results: coding, agents and reasoning. Nemotron 3 Ultra 550B A55B wins following instructions.
Which is better for coding?
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16. It wins 2 of the 5 coding tests both models report; Nemotron 3 Ultra 550B A55B wins none, and 3 are ties.
How do you compare the two?
We use the 22 benchmark tests both models have published scores on. The verdict counts the 9 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.