Model comparison

Nemotron 4 340b Instruct vs Phi 3 Mini 4k Instruct

Nemotron 4 340b Instruct is the stronger model overall, scoring 35.9 to 27.9 on the Noometry Index.

Last verified . 17 shared benchmarks.

Nemotron 4 340b Instruct NVIDIA

35.9

Rank #223 Confirmed

Phi 3 Mini 4k Instruct Microsoft

27.9

Rank #328 Confirmed

Summary

  • They share 17 benchmarks with published results for both. Nemotron 4 340b Instruct scores higher in 8 categories and Phi 3 Mini 4k Instruct in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in instruction following, where Nemotron 4 340b Instruct leads 63.1 to 47.7.

Side by side

Nemotron 4 340b Instruct and Phi 3 Mini 4k Instruct specifications
Nemotron 4 340b InstructPhi 3 Mini 4k Instruct
ProviderNVIDIAMicrosoft
Noometry Index35.927.9
Released2024-06-142024-04-23
WeightsOpenOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked1735

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 35.2 (#228), Phi 3 Mini 4k Instruct: 26.6 (#323)

Coding benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Coding12091093
LiveBench Coding—15.5%
HumanEval+—59.1%
MBPP+—54.2%

Reasoning Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 23.4 (#178), Phi 3 Mini 4k Instruct: 14.1 (#328)

Reasoning benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Hard Prompts12021072
Chess Puzzles—0%
LiveBench Reasoning—26.8%
LiveBench Data Analysis—34.7%
Adversarial NLI—52.8%
BIG-Bench Hard—71.7%
HellaSwag—76.7%
LiveBench—22.4%
WinoGrande—70.8%

Math Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 34.3 (#195), Phi 3 Mini 4k Instruct: 26.6 (#257)

Math benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Math12161111
LiveBench Math—15.7%

Knowledge Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 32.0 (#216), Phi 3 Mini 4k Instruct: 28.5 (#246)

Knowledge benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Expert11711045
ARC (AI2) Challenge—84.9%
MMLU—68.8%
OpenBookQA—88%
TriviaQA—64%

Multilingual Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 37.9 (#233), Phi 3 Mini 4k Instruct: 26.3 (#280)

Multilingual benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Non-English12061021
LMArena Chinese12141021
LMArena French11971076
LMArena German11921044
LMArena Japanese1128935
LMArena Korean1148905
LMArena Russian12191022
LMArena Spanish11951085

Instruction Following Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 63.1 (#232), Phi 3 Mini 4k Instruct: 47.7 (#303)

Instruction Following benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Instruction Following12031053
LiveBench Instruction Following—39.1%

Long Context Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 37.1 (#221), Phi 3 Mini 4k Instruct: 31.7 (#276)

Long Context benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Longer Query12251044

Writing & Preference Nemotron 4 340b Instruct leads

Nemotron 4 340b Instruct: 42.7 (#233), Phi 3 Mini 4k Instruct: 27.6 (#300)

Writing & Preference benchmarks
BenchmarkNemotron 4 340b InstructPhi 3 Mini 4k Instruct
LMArena Text12251073
LMArena Creative Writing12031037
LMArena Multi-Turn12141018
LiveBench Language—9.2%

Frequently asked questions

Is Nemotron 4 340b Instruct better than Phi 3 Mini 4k Instruct?

Nemotron 4 340b Instruct is the stronger model overall, scoring 35.9 to 27.9 on the Noometry Index.

Is Nemotron 4 340b Instruct or Phi 3 Mini 4k Instruct better for coding?

Nemotron 4 340b Instruct scores higher on coding benchmarks: 35.2 versus 26.6 in the Noometry coding category.

How many benchmarks do Nemotron 4 340b Instruct and Phi 3 Mini 4k Instruct share?

17 benchmarks have published results for both models. Nemotron 4 340b Instruct has 17 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.

Related comparisons

Go deeper