DeepSeek-V3 vs NVIDIA Nemotron 3 Nano 30B A3B
Wins 3 of 6 areas
Coding · Facts · Long documents
Wins 3 of 6 areas
Agents · Reasoning · Following instructions
The two are evenly matched.DeepSeek-V3 is better at coding and facts; NVIDIA Nemotron 3 Nano 30B A3B at reasoning and following instructions.
Scores updated · 37 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software41DeepSeek-V34 of 5 tests
- FactsGetting facts right instead of making them up21DeepSeek-V32 of 3 tests
- Long documentsFinding answers in very long texts10DeepSeek-V31 of 1 test
- ReasoningHard problems that need careful thinking02NVIDIA Nemotron 3 Nano 30B A3B2 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked02NVIDIA Nemotron 3 Nano 30B A3B2 of 2 tests
- AgentsCarrying out multi-step tasks on its own01NVIDIA Nemotron 3 Nano 30B A3B1 of 3 tests · 2 ties
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.
NVIDIA Nemotron 3 Nano 30B A3B costs 78% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where DeepSeek-V3 pulls ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+10.2points ahead
- Code for real scientific research problemsSciCode+8.4points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+8.2points ahead
Where NVIDIA Nemotron 3 Nano 30B A3B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+30.1points ahead
- Recent programming contest problemsLiveCodeBench v6+21.4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+10.2points ahead
Every test, side by side
All 37 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V3
- LiveCodeBench v6NVIDIA Nemotron 3 Nano 30B A3B by 21.446.968.3+21.4
- Terminal-Bench 2.1DeepSeek-V3 by 10.216.96.7+10.2
- SciCodeDeepSeek-V3 by 8.43930.6+8.4
- SWE-bench VerifiedDeepSeek-V3 by 3.24238.8+3.2
- Terminal-Bench HardDeepSeek-V3 by 1.615.213.6+1.6
AgentsNVIDIA Nemotron 3 Nano 30B A3B
- τ-Bench V3 · BankingNVIDIA Nemotron 3 Nano 30B A3B by 1.34.76+1.3
- GDPValtie00tie
- Terminal-Bench 4.0tie00tie
ReasoningNVIDIA Nemotron 3 Nano 30B A3B
- GPQA DiamondNVIDIA Nemotron 3 Nano 30B A3B by 10.265.575.7+10.2
- Humanity's Last ExamNVIDIA Nemotron 3 Nano 30B A3B by 6.74.711.4+6.7
- CritPttie00.9tie
FactsDeepSeek-V3
- AA-Omniscience · AccuracyDeepSeek-V3 by 8.225.417.3+8.2
- Vectara HHEM hallucination ratelower is betterDeepSeek-V3 by 3.56.19.6+3.5
- AA-Omniscience · Non-hallucinationNVIDIA Nemotron 3 Nano 30B A3B by 2.614.116.7+2.6
Following instructionsNVIDIA Nemotron 3 Nano 30B A3B
- IFBenchNVIDIA Nemotron 3 Nano 30B A3B by 30.14171.1+30.1
- Multi-ChallengeNVIDIA Nemotron 3 Nano 30B A3B by 7.131.438.5+7.1
Other results20 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.
- OpenBookQADeepSeek-V3 by 48.695.446.8+48.6
- AIME 2025NVIDIA Nemotron 3 Nano 30B A3B by 47.951.399.2+47.9
- τ²-Bench (Retail)DeepSeek-V3 by 12.269.156.9+12.2
- MMLU-ProXDeepSeek-V3 by 1170.559.5+11
- AA-OmniscienceDeepSeek-V3 by 10.9-40.7-51.6+10.9
- MMLUDeepSeek-V3 by 1088.578.5+10
- HumanEvalNVIDIA Nemotron 3 Nano 30B A3B by 9.865.275+9.8
- Tau2 airlineNVIDIA Nemotron 3 Nano 30B A3B by 93948+9
- Artificial Analysis Coding IndexDeepSeek-V3 by 8.62314.4+8.6
- τ²-Bench Telecom (AA run)DeepSeek-V3 by 6.247.140.9+6.2
- WinoGrandeDeepSeek-V3 by 5.784.979.2+5.7
- vectara_factual_consistencyDeepSeek-V3 by 3.593.990.4+3.5
- ARC-ChallengeDeepSeek-V3 by 3.395.392+3.3
- HellaSwagDeepSeek-V3 by 3.388.985.5+3.3
- MMLU-ProNVIDIA Nemotron 3 Nano 30B A3B by 2.475.978.3+2.4
- vectara_avg_summary_lengthNVIDIA Nemotron 3 Nano 30B A3B by 22.5 rating points81.7104.2+22.5 rating
- PIQA (Acc.)DeepSeek-V3 by 2.184.782.6+2.1
- vectara_answer_rateNVIDIA Nemotron 3 Nano 30B A3B by 2.197.599.6+2.1
- MBPPDeepSeek-V3 by 1.675.473.8+1.6
- AA Intelligencetie9.78.9tie
Questions people ask
Which is better, DeepSeek-V3 or NVIDIA Nemotron 3 Nano 30B A3B?
DeepSeek-V3 and NVIDIA Nemotron 3 Nano 30B A3B each win three of the six areas where both have results. DeepSeek-V3 wins coding, facts and long documents; NVIDIA Nemotron 3 Nano 30B A3B wins agents, reasoning and following instructions. NVIDIA Nemotron 3 Nano 30B A3B costs 78% less.
Which is better for coding?
DeepSeek-V3. It wins 4 of the 5 coding tests both models report; NVIDIA Nemotron 3 Nano 30B A3B wins 1.
Which is cheaper?
DeepSeek-V3 costs $0.24 per million input tokens and $0.90 per million output tokens; NVIDIA Nemotron 3 Nano 30B A3B costs $0.05 and $0.20. That makes NVIDIA Nemotron 3 Nano 30B A3B about 78% cheaper for the same work.
How do you compare the two?
We use the 37 benchmark tests both models have published scores on. The verdict counts the 17 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.