Nemotron 3 Super 120B A12B vs Sarvam 30B
Wins 5 of 6 areas
Coding · Reasoning · Facts · Long documents · Following instructions
Wins 1 of 6 areas
Agents
Nemotron 3 Super 120B A12B is the stronger all-rounder.Sarvam 30B is cheaper and better at agents.
Scores updated · 23 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 software30Nemotron 3 Super 120B A12B3 of 3 tests
- ReasoningHard problems that need careful thinking30Nemotron 3 Super 120B A12B3 of 3 tests
- FactsGetting facts right instead of making them up20Nemotron 3 Super 120B A12B2 of 2 tests
- Long documentsFinding answers in very long texts10Nemotron 3 Super 120B A12B1 of 1 test
- Following instructionsDoing exactly what it is asked10Nemotron 3 Super 120B A12B1 of 1 test
- AgentsCarrying out multi-step tasks on its own01Sarvam 30B1 of 2 tests · 1 tie
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.
Sarvam 30B costs 89% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Nemotron 3 Super 120B A12B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+45points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+26.5points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+19.7points ahead
Where Sarvam 30B pulls ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+4.2points ahead
Every test, side by side
All 23 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingNemotron 3 Super 120B A12B
- Terminal-Bench HardNemotron 3 Super 120B A12B by 26.528.82.3+26.5
- SWE-bench VerifiedNemotron 3 Super 120B A12B by 19.753.734+19.7
- SciCodeNemotron 3 Super 120B A12B by 1736.219.2+17
ReasoningNemotron 3 Super 120B A12B
- GPQA DiamondNemotron 3 Super 120B A12B by 16.78063.3+16.7
- Humanity's Last ExamNemotron 3 Super 120B A12B by 13.320.87.5+13.3
- CritPtNemotron 3 Super 120B A12B by 2.83.10.3+2.8
FactsNemotron 3 Super 120B A12B
- AA-Omniscience · AccuracyNemotron 3 Super 120B A12B by 11.724.312.6+11.7
- AA-Omniscience · Non-hallucinationNemotron 3 Super 120B A12B by 9.3133.7+9.3
Long documentsNemotron 3 Super 120B A12B
- AA-LCRNemotron 3 Super 120B A12B by 65.765.70+65.7
Following instructionsNemotron 3 Super 120B A12B
- IFBenchNemotron 3 Super 120B A12B by 4571.526.5+45
Other results11 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)Nemotron 3 Super 120B A12B by 33.367.834.5+33.3
- AA-OmniscienceNemotron 3 Super 120B A12B by 30-41.5-71.5+30
- Artificial Analysis Coding IndexNemotron 3 Super 120B A12B by 29.837.77.9+29.8
- Arena-Hard-V2Nemotron 3 Super 120B A12B by 24.973.949+24.9
- HMMT 2025Nemotron 3 Super 120B A12B by 20.594.774.2+20.5
- HumanEvalSarvam 30B by 11.680.592.1+11.6
- MBPPSarvam 30B by 1181.792.7+11
- AA Agentic IndexSarvam 30B by 7.44.111.5+7.4
- AIME 2025Sarvam 30B by 6.590.296.7+6.5
- AA IntelligenceNemotron 3 Super 120B A12B by 6.212.86.6+6.2
- MMLU-ProNemotron 3 Super 120B A12B by 3.783.780+3.7
Questions people ask
Which is better, Nemotron 3 Super 120B A12B or Sarvam 30B?
Nemotron 3 Super 120B A12B wins five of the six areas where both have results: coding, reasoning, facts, long documents and following instructions. Sarvam 30B wins agents, and costs 89% less.
Which is better for coding?
Nemotron 3 Super 120B A12B. It wins 3 of the 3 coding tests both models report; Sarvam 30B wins none.
Which is cheaper?
Nemotron 3 Super 120B A12B costs $0.30 per million input tokens and $0.90 per million output tokens; Sarvam 30B costs $0.03 and $0.11. That makes Sarvam 30B about 89% cheaper for the same work.
How do you compare the two?
We use the 23 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 11 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.