Enter opens the highlighted result. Esc closes.

Nemotron 3 Ultra vs Sonar

Nemotron 3 Ultra leads Sonar on the Artificial Analysis Intelligence Index, 22.9 to 7.7, and Sonar costs 2.4x less per output token ($1.00 vs $2.40 per million). Updated October 22, 2026; 3 benchmarks have a result for both.

Nemotron 3 UltraNVIDIA
Coding
30.0
Overall
53.1
Output $/M
$2.40
3
SonarPerplexity
Coding
No data
Overall
No data
Output $/M
$1.00
0
vs boards won of 3 shared

Nemotron 3 Ultra or Sonar: which to pick

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Sonar has none, so the two cannot be ranked on coding.

Pick Sonar for output tokens 2.4x cheaper ($1.00 vs $2.40 per million).

Shared results
3
Nemotron 3 Ultra only
68
Sonar only
2
Like-for-like categories
2 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Nemotron 3 Ultra 30.0 from 3 benchmarks, Sonar not benchmarked.
Widest category
Reasoning, 31.6 points apart on average over 2 shared benchmarks.
Widest single gap
GPQA Diamond (Artificial Analysis run): Nemotron 3 Ultra 86.7% against Sonar 47.1%, 39.6 points apart on a 0 to 100 scale.
Output price
Nemotron 3 Ultra $2.40 against Sonar $1.00 per million output tokens, 2.4x apart.
Context window
Nemotron 3 Ultra takes 262K tokens, Sonar 127K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryNemotron 3 UltraBarsSonarTies
Intelligence10
Reasoning20
All shared boards30
CodingDirectional onlySix of these boards feed the Vibe Coding Index14 boards, none shared
Coding benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
AA Coding Index49.3Not measuredOnly Nemotron 3 Ultra measured
LiveCodeBench (Vals AI)86.0%Not measuredOnly Nemotron 3 Ultra measured
SWE-bench Verified (Vals AI)69.0%Not measuredOnly Nemotron 3 Ultra measured
Vibe Code Bench (Vals AI)7.6%Not measuredOnly Nemotron 3 Ultra measured
Code migration (Vals AI)4.9%Not measuredOnly Nemotron 3 Ultra measured
ProgramBench (Vals AI)0.0%Not measuredOnly Nemotron 3 Ultra measured
Terminal-Bench 2.1 (Vals AI)50.9%Not measuredOnly Nemotron 3 Ultra measured
LiveCodeBench (Artificial Analysis run)Not measured29.5%Only Sonar measured
SciCode40.3%Not measuredOnly Nemotron 3 Ultra measured
Terminal-Bench Hard (Artificial Analysis run)36.4%Not measuredOnly Nemotron 3 Ultra measured
FrontierCode13.6%Not measuredOnly Nemotron 3 Ultra measured
WeirdML43.5%Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: coding1468Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: software and IT1468Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 29 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
AA Intelligence Index22.97.7Nemotron 3 Ultra
Epoch Capabilities Index146.0Not measuredOnly Nemotron 3 Ultra measured
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
Agentic work benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
AA Agentic Index20.1Not measuredOnly Nemotron 3 Ultra measured
Finance Agent (Vals AI)37.7%Not measuredOnly Nemotron 3 Ultra measured
Legal Research Bench (Vals AI)15.4%Not measuredOnly Nemotron 3 Ultra measured
Excel Modeling Benchmark (Vals AI)31.7%Not measuredOnly Nemotron 3 Ultra measured
τ²-Bench Telecom83.3%Not measuredOnly Nemotron 3 Ultra measured
APEX-Agents22.7%Not measuredOnly Nemotron 3 Ultra measured
MCP Atlas63.1%Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 35 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
Reasoning benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
GPQA Diamond86.1%Not measuredOnly Nemotron 3 Ultra measured
GPQA Diamond (Artificial Analysis run)86.7%47.1%Nemotron 3 Ultra
Humanity's Last Exam (Artificial Analysis run)28.4%4.9%Nemotron 3 Ultra
GPQA Diamond (Epoch AI run)85.4%Not measuredOnly Nemotron 3 Ultra measured
Mystery game puzzles20.0%Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: hard prompts1452Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: expert1472Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: hard prompts in English1453Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 15 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
Math benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
OTIS Mock AIME86.7%Not measuredOnly Nemotron 3 Ultra measured
ProofBench2.0%Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: math1457Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: mathematical work1461Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 10 boards here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index10 boards, none shared
Knowledge benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
MMLU Pro85.8%Not measuredOnly Nemotron 3 Ultra measured
LegalBench (Vals AI)82.1%Not measuredOnly Nemotron 3 Ultra measured
CorpFin (Vals AI)65.5%Not measuredOnly Nemotron 3 Ultra measured
TaxEval (Vals AI)73.1%Not measuredOnly Nemotron 3 Ultra measured
MedCode (Vals AI)38.6%Not measuredOnly Nemotron 3 Ultra measured
MMLU Pro (Artificial Analysis run)Not measured68.9%Only Sonar measured
LMArena Text: science1465Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: medicine and healthcare1456Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: legal and government1452Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: business and finance1434Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 10 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
Design Arena: websites1145Not measuredOnly Nemotron 3 Ultra measured
Design Arena: UI components1135Not measuredOnly Nemotron 3 Ultra measured
Design Arena: game dev1138Not measuredOnly Nemotron 3 Ultra measured
Design Arena: data viz1141Not measuredOnly Nemotron 3 Ultra measured
Design Arena: 3D1131Not measuredOnly Nemotron 3 Ultra measured
Design Arena: all code categories1146Not measuredOnly Nemotron 3 Ultra measured
Design Arena: SVG1069Not measuredOnly Nemotron 3 Ultra measured
Design Arena: ASCII art1100Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 1 board here
GeneralDirectional onlyPublic boards, shown as published, kept out of the index20 boards, none shared
General benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
IFBench81.4%Not measuredOnly Nemotron 3 Ultra measured
Long context reasoning (AA-LCR)79.3%Not measuredOnly Nemotron 3 Ultra measured
DTBench90.1%Not measuredOnly Nemotron 3 Ultra measured
LMCA36.9%Not measuredOnly Nemotron 3 Ultra measured
LMArena Text1445Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: instruction following1418Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: creative writing1399Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: English1455Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: non-English1427Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Chinese1497Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: German1472Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Korean1384Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: French1472Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Spanish1455Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Russian1417Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Polish1385Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: multi-turn1411Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: longer queries1437Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: writing and language1406Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: entertainment and media1397Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 28 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Nemotron 3 Ultra and Sonar
BenchmarkNemotron 3 UltraSonarLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Nemotron 3 Ultra$0.0018Fits in one request
Sonar$0.0015Fits in one request

Sonar costs less

Repository review

50K input + 3K output tokens

Nemotron 3 Ultra$0.037Fits in one request
Sonar$0.053Fits in one request

Nemotron 3 Ultra costs less

Agent loop

200K input + 10K output tokens

Nemotron 3 Ultra$0.144Fits in one request
Sonar$0.210Exceeds the context window

Nemotron 3 Ultra costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Nemotron 3 Ultra and Sonar
SpecificationNVIDIA Nemotron 3 UltraPerplexity Sonar
MakerNVIDIAPerplexity
WeightsOpen weightsNo public weights listed
ReleasedJune 4, 2026January 27, 2025
Context window262K127K
Input $/Mtok$0.60$1.00
Output $/Mtok$2.40$1.00
Coding index30.0Not benchmarked
Independent measurements behind it30
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Nemotron 3 Ultra better than Sonar for coding?

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Sonar has none, so the two cannot be ranked on coding. Pick Sonar for output tokens 2.4x cheaper ($1.00 vs $2.40 per million).

Which is cheaper, Nemotron 3 Ultra or Sonar?

Nemotron 3 Ultra costs $0.60 per million input tokens and $2.40 per million output tokens; Sonar costs $1.00 and $1.00. Cost per solved task is known for neither of them.

Which has the bigger context window, Nemotron 3 Ultra or Sonar?

Nemotron 3 Ultra, with 262K tokens against 127K.

How many benchmarks is the Nemotron 3 Ultra vs Sonar comparison based on?

3 benchmarks have a result for both models as of October 22, 2026. Nemotron 3 Ultra's coding index rests on 3 independent benchmarks and Sonar's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next