Nemotron 3 Super 120B A12B vs Qwen3.5 35B A3B
Wins 0 of 7 areas
—
Wins 4 of 7 areas
Coding · Agents · Long documents · Following instructions
Qwen3.5 35B A3B wins more areas, narrowly.Nemotron 3 Super 120B A12B is cheaper.
Scores updated · 40 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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 software14Qwen3.5 35B A3B4 of 5 tests
- Following instructionsDoing exactly what it is asked02Qwen3.5 35B A3B2 of 2 tests
- AgentsCarrying out multi-step tasks on its own12Qwen3.5 35B A3B2 of 3 tests
- Long documentsFinding answers in very long texts01Qwen3.5 35B A3B1 of 1 test
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
- FactsGetting facts right instead of making them up11Even1 each
- MathCompetition and research-level math11Even1 each
Long documents rests on a single test.
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.
Nemotron 3 Super 120B A12B costs 47% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.5 35B A3B pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+29.7points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+15.5points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+14.5points ahead
Where Nemotron 3 Super 120B A12B pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+5.4points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+4.2points ahead
- Harvard-MIT high-school math contest problemsHMMT Feb. 2026+3.1points ahead
Every test, side by side
All 40 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.5 35B A3B
- SWE-bench VerifiedQwen3.5 35B A3B by 15.553.769.2+15.5
- SWE-bench MultilingualQwen3.5 35B A3B by 14.545.860.3+14.5
- Terminal-Bench HardNemotron 3 Super 120B A12B by 2.328.826.5+2.3
- Terminal-Bench 2.1Qwen3.5 35B A3B by 2.238.640.8+2.2
- SciCodeQwen3.5 35B A3B by 1.536.237.7+1.5
AgentsQwen3.5 35B A3B
- BrowseCompQwen3.5 35B A3B by 29.731.361+29.7
- GDPValQwen3.5 35B A3B by 5.605.6+5.6
- τ-Bench V3 · BankingNemotron 3 Super 120B A12B by 5.410.34.9+5.4
ReasoningEven
- GPQA DiamondQwen3.5 35B A3B by 4.58084.5+4.5
- CritPtNemotron 3 Super 120B A12B by 2.23.10.9+2.2
- Humanity's Last Examtie20.821tie
FactsEven
- AA-Omniscience · AccuracyNemotron 3 Super 120B A12B by 4.224.320.1+4.2
- AA-Omniscience · Non-hallucinationQwen3.5 35B A3B by 1.61314.6+1.6
MathEven
- HMMT Feb. 2026Nemotron 3 Super 120B A12B by 3.184.881.8+3.1
- AIME 2026Qwen3.5 35B A3B by 1.691.793.3+1.6
Following instructionsQwen3.5 35B A3B
- Multi-ChallengeQwen3.5 35B A3B by 4.855.260+4.8
- IFBenchQwen3.5 35B A3B by 171.572.5+1
Other results22 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.
- RULER 1MNemotron 3 Super 120B A12B by 37.593.956.4+37.5
- Minerva MathNemotron 3 Super 120B A12B by 25.184.859.7+25.1
- τ²-Bench Telecom (AA run)Qwen3.5 35B A3B by 21.467.889.2+21.4
- RULER 256KNemotron 3 Super 120B A12B by 14.396.782.4+14.3
- HumanEvalNemotron 3 Super 120B A12B by 1480.566.5+14
- MBPPNemotron 3 Super 120B A12B by 10.981.770.8+10.9
- Terminal-Bench 2.0Qwen3.5 35B A3B by 9.53140.5+9.5
- AA Agentic IndexQwen3.5 35B A3B by 7.74.111.8+7.7
- AA-OmniscienceNemotron 3 Super 120B A12B by 6.6-41.5-48.1+6.6
- AA IntelligenceQwen3.5 35B A3B by 6.512.819.3+6.5
- HMMT 2025Nemotron 3 Super 120B A12B by 5.594.789.2+5.5
- MMLUNemotron 3 Super 120B A12B by 4.98681.1+4.9
- Global-MMLU-Lite (5-shot) AverageNemotron 3 Super 120B A12B by 4.885.780.9+4.8
- OpenBookQANemotron 3 Super 120B A12B by 4.448.644.2+4.4
- HellaSwagNemotron 3 Super 120B A12B by 3.48985.6+3.4
- PIQA (acc-norm)Nemotron 3 Super 120B A12B by 385.582.5+3
- MMLU-ProQwen3.5 35B A3B by 1.683.785.3+1.6
- MMLU-ProXQwen3.5 35B A3B by 1.679.481+1.6
- PIQA (Acc.)Nemotron 3 Super 120B A12B by 1.683.982.3+1.6
- ARC-Challengetie96.195.4tie
- Artificial Analysis Coding Indextie37.737tie
- WinoGrandetie78.979.2tie
Questions people ask
Which is better, Nemotron 3 Super 120B A12B or Qwen3.5 35B A3B?
Qwen3.5 35B A3B wins four of the seven areas where both have results: coding, agents, long documents and following instructions. Nemotron 3 Super 120B A12B wins none, but costs 47% less. They are level on reasoning, facts and math.
Which is better for coding?
Qwen3.5 35B A3B. It wins 4 of the 5 coding tests both models report; Nemotron 3 Super 120B A12B wins 1.
Which is cheaper?
Nemotron 3 Super 120B A12B costs $0.30 per million input tokens and $0.90 per million output tokens; Qwen3.5 35B A3B costs $0.25 and $2.00. That makes Nemotron 3 Super 120B A12B about 47% cheaper for the same work.
How do you compare the two?
We use the 40 benchmark tests both models have published scores on. The verdict counts the 18 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 22 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.