NVIDIA Nemotron 3 Nano 30B A3B vs Phi 4 Mini Instruct
Wins 6 of 6 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 0 of 6 areas
—
NVIDIA Nemotron 3 Nano 30B A3B is the stronger all-rounder.
Scores updated · 24 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.
- ReasoningHard problems that need careful thinking20NVIDIA Nemotron 3 Nano 30B A3B2 of 3 tests · 1 tie
- CodingWriting and fixing software20NVIDIA Nemotron 3 Nano 30B A3B2 of 2 tests
- FactsGetting facts right instead of making them up21NVIDIA Nemotron 3 Nano 30B A3B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own10NVIDIA Nemotron 3 Nano 30B A3B1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10NVIDIA Nemotron 3 Nano 30B A3B1 of 1 test
- Following instructionsDoing exactly what it is asked10NVIDIA Nemotron 3 Nano 30B A3B1 of 1 test
Long documents 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 Nano 30B A3B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+50points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+42.6points ahead
- Reasons across sets of long documentsAA-LCR+22.7points ahead
Where Phi 4 Mini Instruct pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+5.3points ahead
Every test, side by side
All 24 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingNVIDIA Nemotron 3 Nano 30B A3B
- SciCodeNVIDIA Nemotron 3 Nano 30B A3B by 19.830.610.8+19.8
- Terminal-Bench HardNVIDIA Nemotron 3 Nano 30B A3B by 13.613.60+13.6
AgentsNVIDIA Nemotron 3 Nano 30B A3B
- τ-Bench V3 · BankingNVIDIA Nemotron 3 Nano 30B A3B by 561+5
- GDPValtie00tie
ReasoningNVIDIA Nemotron 3 Nano 30B A3B
- GPQA DiamondNVIDIA Nemotron 3 Nano 30B A3B by 42.675.733.1+42.6
- Humanity's Last ExamNVIDIA Nemotron 3 Nano 30B A3B by 6.911.44.5+6.9
- CritPttie0.90tie
FactsNVIDIA Nemotron 3 Nano 30B A3B
- Vectara HHEM hallucination ratelower is betterNVIDIA Nemotron 3 Nano 30B A3B by 13.99.623.5+13.9
- AA-Omniscience · AccuracyNVIDIA Nemotron 3 Nano 30B A3B by 7.817.39.5+7.8
- AA-Omniscience · Non-hallucinationPhi 4 Mini Instruct by 5.316.722+5.3
Long documentsNVIDIA Nemotron 3 Nano 30B A3B
- AA-LCRNVIDIA Nemotron 3 Nano 30B A3B by 22.73815.3+22.7
Following instructionsNVIDIA Nemotron 3 Nano 30B A3B
- IFBenchNVIDIA Nemotron 3 Nano 30B A3B by 5071.121.1+50
Other results12 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.
- τ²-Bench Telecom (AA run)NVIDIA Nemotron 3 Nano 30B A3B by 32.740.98.2+32.7
- OpenBookQAPhi 4 Mini Instruct by 32.446.879.2+32.4
- vectara_avg_summary_lengthPhi 4 Mini Instruct by 316 rating points104.2420.2+316 rating
- MMLU-ProNVIDIA Nemotron 3 Nano 30B A3B by 25.578.352.8+25.5
- HellaSwagNVIDIA Nemotron 3 Nano 30B A3B by 16.585.569.1+16.5
- vectara_factual_consistencyNVIDIA Nemotron 3 Nano 30B A3B by 13.990.476.5+13.9
- WinoGrandeNVIDIA Nemotron 3 Nano 30B A3B by 12.279.267+12.2
- Artificial Analysis Coding IndexNVIDIA Nemotron 3 Nano 30B A3B by 10.614.43.8+10.6
- AA-OmniscienceNVIDIA Nemotron 3 Nano 30B A3B by 9.5-51.6-61.1+9.5
- ARC-ChallengeNVIDIA Nemotron 3 Nano 30B A3B by 8.39283.7+8.3
- vectara_answer_rateNVIDIA Nemotron 3 Nano 30B A3B by 7.199.692.5+7.1
- AA IntelligenceNVIDIA Nemotron 3 Nano 30B A3B by 2.68.96.3+2.6
Questions people ask
Which is better, NVIDIA Nemotron 3 Nano 30B A3B or Phi 4 Mini Instruct?
NVIDIA Nemotron 3 Nano 30B A3B wins all six areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Phi 4 Mini Instruct wins none.
Which is better for coding?
NVIDIA Nemotron 3 Nano 30B A3B. It wins 2 of the 2 coding tests both models report; Phi 4 Mini Instruct wins none.
How do you compare the two?
We use the 24 benchmark tests both models have published scores on. The verdict counts the 12 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 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.