Model comparison

Gemma 4 31B IT vs Gemma 7B

Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 30.0 on the Noometry Index.

Last verified . 14 shared benchmarks.

Gemma 4 31B IT Google

43.5

Rank #90 Confirmed

Gemma 7B Google

30.0

Rank #299 Confirmed

Summary

  • They share 14 benchmarks with published results for both. Gemma 4 31B IT scores higher in 8 categories and Gemma 7B in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Gemma 4 31B IT leads 60.5 to 27.1.

Side by side

Gemma 4 31B IT and Gemma 7B specifications
Gemma 4 31B ITGemma 7B
ProviderGoogleGoogle
Noometry Index43.530.0
Released2026-04-022024-02-21
WeightsOpenOpen
Context window262K—
Max output33K—
Input $ / M tokens$0.09—
Output $ / M tokens$0.34—
Results tracked3527

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

Category by category

Coding Gemma 4 31B IT leads

Gemma 4 31B IT: 42.3 (#108), Gemma 7B: 30.5 (#294)

Coding benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Coding14591048
LMArena WebDev1366—
SciCode43.4%—
WeirdML52.3%—
ALE-Bench925.5—
HumanEval+—28.7%
MBPP+—43.4%

Reasoning Gemma 4 31B IT leads

Gemma 4 31B IT: 27.2 (#122), Gemma 7B: 19.9 (#249)

Reasoning benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Hard Prompts14481042
Epoch Capabilities Index142.74111.99
Kagi LLM Benchmark63.5%—
NYT Connections (extended)70.6%—
CritPt1.4%—
Chess Puzzles5%—
Thematic Generalization53%—
DTBench82.7%—
LMCA39.3%—
Surface Evolver Bench30.6%—
Adversarial NLI—48.7%
BIG-Bench Hard—55.1%
HellaSwag—82.2%
PIQA—81.2%
WinoGrande—79%

Math Gemma 4 31B IT leads

Gemma 4 31B IT: 43.2 (#81), Gemma 7B: 31.2 (#228)

Math benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Math14651066
OTIS Mock AIME 2024-202573.3%—
GSM8K—46.4%

Knowledge Gemma 4 31B IT leads

Gemma 4 31B IT: 37.9 (#151), Gemma 7B: 27.3 (#252)

Knowledge benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Expert14651001
GPQA Diamond75.8%—
SimpleQA Verified10.4%—
Vectara Hallucination Rate7.4%—
ARC (AI2) Challenge—78.3%
BoolQ—83.2%
MMLU—66.1%
OpenBookQA—78.6%
TriviaQA—72.3%

Multimodal Not comparable

Gemma 4 31B IT: 41.6 (#34), Gemma 7B: —

Multimodal benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Vision1277—
LMArena Document1425—

Multilingual Gemma 4 31B IT leads

Gemma 4 31B IT: 53.8 (#57), Gemma 7B: 25.1 (#287)

Multilingual benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Non-English1431999
LMArena Chinese14761035
LMArena French14351025
LMArena Russian1460993
LMArena Spanish1444—

Instruction Following Gemma 4 31B IT leads

Gemma 4 31B IT: 75.5 (#61), Gemma 7B: 51.5 (#295)

Instruction Following benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Instruction Following14331017

Long Context Gemma 4 31B IT leads

Gemma 4 31B IT: 44.2 (#71), Gemma 7B: 31.1 (#282)

Long Context benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Longer Query14461022

Writing & Preference Gemma 4 31B IT leads

Gemma 4 31B IT: 60.5 (#96), Gemma 7B: 27.1 (#302)

Writing & Preference benchmarks
BenchmarkGemma 4 31B ITGemma 7B
LMArena Text14431056
LMArena Creative Writing14151024
LMArena Multi-Turn1452963
EQ-Bench Creative Writing1368—
EQ-Bench 41120—

Frequently asked questions

Is Gemma 4 31B IT better than Gemma 7B?

Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 30.0 on the Noometry Index.

Is Gemma 4 31B IT or Gemma 7B better for coding?

Gemma 4 31B IT scores higher on coding benchmarks: 42.3 versus 30.5 in the Noometry coding category.

How many benchmarks do Gemma 4 31B IT and Gemma 7B share?

14 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Gemma 7B has 27.

Related comparisons

Go deeper