Model comparison

Gemma 3 27B vs Step 3

Step 3 is the stronger model overall, scoring 39.9 to 30.5 on the Noometry Index.

Last verified . 17 shared benchmarks.

Gemma 3 27B Google

30.5

Rank #277 Confirmed

Step 3 StepFun

39.9

Rank #145 Confirmed

Summary

  • They share 17 benchmarks with published results for both. Gemma 3 27B scores higher in 2 categories and Step 3 in 7 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in coding, where Step 3 leads 39.5 to 22.1.
  • The biggest single-benchmark swing is Kagi LLM Benchmark: 40.4% for Gemma 3 27B and 62.3% for Step 3.

Side by side

Gemma 3 27B and Step 3 specifications
Gemma 3 27BStep 3
ProviderGoogleStepFun
Noometry Index30.539.9
Released2025-03-11—
WeightsOpenOpen
Context window131K—
Max output8K—
Input $ / M tokens$0.08—
Output $ / M tokens$0.16—
Results tracked4317

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

Category by category

Coding Step 3 leads

Gemma 3 27B: 22.1 (#329), Step 3: 39.5 (#144)

Coding benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Coding13221367
Aider Polyglot4.9%—
SciCode21.2%—
LiveBench Coding39.9%—

Agentic & Tool Use Not comparable

Gemma 3 27B: 24.9 (#105), Step 3: —

Agentic & Tool Use benchmarks
BenchmarkGemma 3 27BStep 3
Berkeley Function Calling Leaderboard29.5%—

Reasoning Step 3 leads

Gemma 3 27B: 16.3 (#297), Step 3: 27.6 (#102)

Reasoning benchmarks
BenchmarkGemma 3 27BStep 3
Kagi LLM Benchmark40.4%62.3%
LMArena Hard Prompts13401355
CritPt0%—
Chess Puzzles0%—
LiveBench Reasoning43.8%—
DTBench52.5%—
LiveBench Data Analysis51.5%—
LMCA12.3%—
Epoch Capabilities Index130.04—
LiveBench50%—

Math Step 3 leads

Gemma 3 27B: 24.4 (#261), Step 3: 34.9 (#148)

Math benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Math13121366
OTIS Mock AIME 2024-202522.5%—
LiveBench Math55.4%—
MATH Level 574%—

Knowledge Step 3 leads

Gemma 3 27B: 25.4 (#256), Step 3: 36.2 (#161)

Knowledge benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Expert13031333
GPQA Diamond47.7%—
Confabulations40.3%—
Vectara Hallucination Rate7.4%—

Multimodal Step 3 leads

Gemma 3 27B: 32.0 (#94), Step 3: 34.9 (#81)

Multimodal benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Vision11651176
GeoBench52%—

Multilingual Too close to call

Gemma 3 27B: 47.0 (#150), Step 3: 46.5 (#154)

Multilingual benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Non-English13341327
LMArena Chinese13471396
LMArena German13621371
LMArena Korean13081271
LMArena Russian13491331
LMArena Spanish13481372
LMArena French1368—
LMArena Japanese1287—

Instruction Following Too close to call

Gemma 3 27B: 70.8 (#155), Step 3: 70.6 (#159)

Instruction Following benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Instruction Following13211332
LiveBench Instruction Following74.9%—

Long Context Step 3 leads

Gemma 3 27B: 27.6 (#288), Step 3: 40.3 (#152)

Long Context benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Longer Query13331326
Fiction.LiveBench33.3%—

Writing & Preference Step 3 leads

Gemma 3 27B: 52.7 (#161), Step 3: 54.5 (#146)

Writing & Preference benchmarks
BenchmarkGemma 3 27BStep 3
LMArena Text13581351
LMArena Creative Writing13461322
LMArena Multi-Turn13441340
Short-Story Creative Writing79.9%—
EQ-Bench Creative Writing1266—
LiveBench Language34.6%—

Frequently asked questions

Is Gemma 3 27B better than Step 3?

Step 3 is the stronger model overall, scoring 39.9 to 30.5 on the Noometry Index.

Is Gemma 3 27B or Step 3 better for coding?

Step 3 scores higher on coding benchmarks: 39.5 versus 22.1 in the Noometry coding category.

How many benchmarks do Gemma 3 27B and Step 3 share?

17 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Step 3 has 17.

Related comparisons

Go deeper