Enter opens the highlighted result. Esc closes.

Phi 4 vs Nemotron 3 Ultra

Nemotron 3 Ultra leads Phi 4 on the Artificial Analysis Intelligence Index, 22.9 to 5.9, and Phi 4 costs 17.1x less per output token ($0.14 vs $2.40 per million). Updated October 22, 2026; 36 benchmarks have a result for both.

Replace Phi 4 Replace Nemotron 3 Ultra
Phi 4Microsoft
Coding
No data
Overall
24.7
Output $/M
$0.14
0
Nemotron 3 UltraNVIDIA
Coding
30.0
Overall
53.1
Output $/M
$2.40
36
vs boards won of 36 shared

Phi 4 or Nemotron 3 Ultra: which to pick

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Phi 4 has none, so the two cannot be ranked on coding.

Pick Phi 4 for output tokens 17.1x cheaper ($0.14 vs $2.40 per million).

Model APhi 4MicrosoftNot benchmarkedNo coding benchmarkNo cost per solved taskModel BNemotron 3 UltraNVIDIA30.0/1003 independent benchmarksNo cost per solved task
Shared results
36
Phi 4 only
10
Nemotron 3 Ultra only
34
Like-for-like categories
7 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
Phi 4 not benchmarked, Nemotron 3 Ultra 30.0 from 3 benchmarks.
Widest category
Agentic work, 83.3 points apart on average over 1 shared benchmark.
Widest single gap
τ²-Bench Telecom: Phi 4 0.0% against Nemotron 3 Ultra 83.3%, 83.3 points apart on a 0 to 100 scale.
Output price
Phi 4 $0.14 against Nemotron 3 Ultra $2.40 per million output tokens, 17.1x apart.
Context window
Phi 4 takes 16K tokens, Nemotron 3 Ultra 262K.

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
CategoryPhi 4BarsNemotron 3 UltraTies
Coding03
Intelligence02
Agentic work01
Reasoning06
Math03
Knowledge04
General017
All shared boards036
CodingLike-for-likeSix of these boards feed the Vibe Coding Index03
Coding benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
AA Coding IndexNot measured49.3Only Nemotron 3 Ultra measured
LiveCodeBench (Vals AI)Not measured86.0%Only Nemotron 3 Ultra measured
SWE-bench Verified (Vals AI)Not measured69.0%Only Nemotron 3 Ultra measured
Vibe Code Bench (Vals AI)Not measured7.6%Only Nemotron 3 Ultra measured
Code migration (Vals AI)Not measured4.9%Only Nemotron 3 Ultra measured
ProgramBench (Vals AI)Not measured0.0%Only Nemotron 3 Ultra measured
Terminal-Bench 2.1 (Vals AI)Not measured50.9%Only Nemotron 3 Ultra measured
LiveCodeBench (Artificial Analysis run)23.1%Not measuredOnly Phi 4 measured
SciCodeNot measured40.3%Only Nemotron 3 Ultra measured
Terminal-Bench Hard (Artificial Analysis run)3.8%36.4%Nemotron 3 Ultra
FrontierCodeNot measured13.6%Only Nemotron 3 Ultra measured
WeirdMLNot measured43.5%Only Nemotron 3 Ultra measured
LMArena Text: coding12311468Nemotron 3 Ultra
LMArena Text: software and IT12281468Nemotron 3 Ultra
Neither model has a result on the other 25 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
AA Intelligence Index5.922.9Nemotron 3 Ultra
Epoch Capabilities Index130.5146.0Nemotron 3 Ultra
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
AA Agentic IndexNot measured20.1Only Nemotron 3 Ultra measured
Finance Agent (Vals AI)Not measured37.7%Only Nemotron 3 Ultra measured
Legal Research Bench (Vals AI)Not measured15.4%Only Nemotron 3 Ultra measured
Excel Modeling Benchmark (Vals AI)Not measured31.7%Only Nemotron 3 Ultra measured
τ²-Bench Telecom0.0%83.3%Nemotron 3 Ultra
APEX-AgentsNot measured22.7%Only Nemotron 3 Ultra measured
BALROG11.6%Not measuredOnly Phi 4 measured
Berkeley Function Calling Leaderboard (BFCL)28.8%Not measuredOnly Phi 4 measured
Neither model has a result on the other 33 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index06
Reasoning benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
GPQA DiamondNot measured86.1%Only Nemotron 3 Ultra measured
GPQA Diamond (Artificial Analysis run)57.5%86.7%Nemotron 3 Ultra
Humanity's Last Exam (Artificial Analysis run)3.8%28.4%Nemotron 3 Ultra
GPQA Diamond (Epoch AI run)56.1%85.4%Nemotron 3 Ultra
Mystery game puzzlesNot measured20.0%Only Nemotron 3 Ultra measured
LMArena Text: hard prompts12201452Nemotron 3 Ultra
LMArena Text: expert12031472Nemotron 3 Ultra
LMArena Text: hard prompts in English12301453Nemotron 3 Ultra
Neither model has a result on the other 14 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index03
Math benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
AA Math Index18.0Not measuredOnly Phi 4 measured
AIME 2025 (Artificial Analysis run)18.0%Not measuredOnly Phi 4 measured
OTIS Mock AIME13.8%86.7%Nemotron 3 Ultra
ProofBenchNot measured2.0%Only Nemotron 3 Ultra measured
MATH level 564.9%Not measuredOnly Phi 4 measured
LMArena Text: math12461457Nemotron 3 Ultra
LMArena Text: mathematical work12541461Nemotron 3 Ultra
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
Knowledge benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
MMLU ProNot measured85.8%Only Nemotron 3 Ultra measured
LegalBench (Vals AI)Not measured82.1%Only Nemotron 3 Ultra measured
CorpFin (Vals AI)Not measured65.5%Only Nemotron 3 Ultra measured
TaxEval (Vals AI)Not measured73.1%Only Nemotron 3 Ultra measured
MedCode (Vals AI)Not measured38.6%Only Nemotron 3 Ultra measured
MMLU Pro (Artificial Analysis run)71.4%Not measuredOnly Phi 4 measured
MMLU84.8%Not measuredOnly Phi 4 measured
LMArena Text: science12061465Nemotron 3 Ultra
LMArena Text: medicine and healthcare12021456Nemotron 3 Ultra
LMArena Text: legal and government12411452Nemotron 3 Ultra
LMArena Text: business and finance11991434Nemotron 3 Ultra
Neither model has a result on the other 7 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
Design Arena: websitesNot measured1145Only Nemotron 3 Ultra measured
Design Arena: UI componentsNot measured1136Only Nemotron 3 Ultra measured
Design Arena: game devNot measured1138Only Nemotron 3 Ultra measured
Design Arena: data vizNot measured1142Only Nemotron 3 Ultra measured
Design Arena: 3DNot measured1131Only Nemotron 3 Ultra measured
Design Arena: all code categoriesNot measured1146Only Nemotron 3 Ultra measured
Design Arena: SVGNot measured1069Only Nemotron 3 Ultra measured
Design Arena: ASCII artNot measured1100Only Nemotron 3 Ultra measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index017
General benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
IFBench23.5%81.4%Nemotron 3 Ultra
Long context reasoning (AA-LCR)0.0%79.3%Nemotron 3 Ultra
DTBenchNot measured90.1%Only Nemotron 3 Ultra measured
LMCANot measured36.9%Only Nemotron 3 Ultra measured
Creative writing (Lech Mazur)62.6%Not measuredOnly Phi 4 measured
LMArena Text12171445Nemotron 3 Ultra
LMArena Text: instruction following12011418Nemotron 3 Ultra
LMArena Text: creative writing11811399Nemotron 3 Ultra
LMArena Text: English12311455Nemotron 3 Ultra
LMArena Text: non-English11971427Nemotron 3 Ultra
LMArena Text: Chinese12121497Nemotron 3 Ultra
LMArena Text: German12221472Nemotron 3 Ultra
LMArena Text: Japanese1158Not measuredOnly Phi 4 measured
LMArena Text: Korean11511384Nemotron 3 Ultra
LMArena Text: French12241472Nemotron 3 Ultra
LMArena Text: Spanish12341455Nemotron 3 Ultra
LMArena Text: Russian12091417Nemotron 3 Ultra
LMArena Text: PolishNot measured1385Only Nemotron 3 Ultra measured
LMArena Text: multi-turn12061411Nemotron 3 Ultra
LMArena Text: longer queries12171437Nemotron 3 Ultra
LMArena Text: writing and language11981406Nemotron 3 Ultra
LMArena Text: entertainment and media11621397Nemotron 3 Ultra
Neither model has a result on the other 24 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Phi 4 and Nemotron 3 Ultra
BenchmarkPhi 4Nemotron 3 UltraLeader
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

Phi 4$0.0001Fits in one request
Nemotron 3 Ultra$0.0018Fits in one request

Phi 4 costs less

Repository review

50K input + 3K output tokens

Phi 4$0.0039Exceeds the context window
Nemotron 3 Ultra$0.037Fits in one request

Phi 4 costs less

Agent loop

200K input + 10K output tokens

Phi 4$0.015Exceeds the context window
Nemotron 3 Ultra$0.144Fits in one request

Phi 4 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 Phi 4 and Nemotron 3 Ultra
SpecificationMicrosoft Phi 4NVIDIA Nemotron 3 Ultra
MakerMicrosoftNVIDIA
WeightsOpen weightsOpen weights
ReleasedJanuary 10, 2025June 4, 2026
Context window16K262K
Input $/Mtok$0.07$0.60
Output $/Mtok$0.14$2.40
Coding indexNot benchmarked30.0
Independent measurements behind it03
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Phi 4 better than Nemotron 3 Ultra for coding?

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Phi 4 has none, so the two cannot be ranked on coding. Pick Phi 4 for output tokens 17.1x cheaper ($0.14 vs $2.40 per million).

Which is cheaper, Phi 4 or Nemotron 3 Ultra?

Phi 4 costs $0.07 per million input tokens and $0.14 per million output tokens; Nemotron 3 Ultra costs $0.60 and $2.40. Cost per solved task is known for neither of them.

Which has the bigger context window, Phi 4 or Nemotron 3 Ultra?

Nemotron 3 Ultra, with 262K tokens against 16K.

How many benchmarks is the Phi 4 vs Nemotron 3 Ultra comparison based on?

36 benchmarks have a result for both models as of October 22, 2026. Phi 4's coding index rests on 0 independent benchmarks and Nemotron 3 Ultra's on 3. Where only one model has a result, the page says so instead of naming a winner.

Compare next