Model comparison

Gemma 2 27B vs StarCoder 2 15B

Gemma 2 27B has enough public results to be ranked (#312); StarCoder 2 15B does not yet, so treat this comparison as directional.

Last verified . 3 shared benchmarks.

Gemma 2 27B Google

29.4

Rank #312 Confirmed

StarCoder 2 15B NVIDIA

38.1

Unranked Sparse

Summary

  • They share 3 benchmarks with published results for both. Gemma 2 27B scores higher in 0 categories and StarCoder 2 15B in 1 category; one gap is clear of the uncertainty.
  • The biggest single-benchmark swing is BigCodeBench Complete: 52.5% for Gemma 2 27B and 38.4% for StarCoder 2 15B.

Side by side

Gemma 2 27B and StarCoder 2 15B specifications
Gemma 2 27BStarCoder 2 15B
ProviderGoogleNVIDIA
Noometry Index29.438.1
Released2024-06-242024-02-20
WeightsOpenOpen
Context window8K—
Max output2K—
Input $ / M tokens$0.65—
Output $ / M tokens$0.65—
Results tracked347

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

Category by category

Coding StarCoder 2 15B leads

Gemma 2 27B: 34.1 (#246), StarCoder 2 15B: 36.6

Coding benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
BigCodeBench Complete52.5%38.4%
BigCodeBench Instruct42.8%—
LiveBench Coding36%—
LMArena Coding1211—
HumanEval+—37.8%

Reasoning Not comparable

Gemma 2 27B: 15.3 (#315), StarCoder 2 15B: —

Reasoning benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
Epoch Capabilities Index122.08105
LiveBench Reasoning28.1%—
LMArena Hard Prompts1198—
DTBench48%—
LiveBench Data Analysis47.9%—
LMCA7.1%—
LiveBench38.2%—
WinoGrande—64.3%

Math Not comparable

Gemma 2 27B: 10.7 (#311), StarCoder 2 15B: —

Math benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
OTIS Mock AIME 2024-20251.4%—
LiveBench Math26.5%—
LMArena Math1212—
MATH Level 527.9%—
GSM8K—57.7%

Knowledge Not comparable

Gemma 2 27B: 19.0 (#280), StarCoder 2 15B: —

Knowledge benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
MMLU75.7%64.1%
GPQA Diamond36.5%—
Confabulations27.1%—
LMArena Expert1172—
ARC (AI2) Challenge—47.2%

Multilingual Not comparable

Gemma 2 27B: 38.6 (#226), StarCoder 2 15B: —

Multilingual benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
LMArena Non-English1217—
LMArena Chinese1221—
LMArena French1247—
LMArena German1209—
LMArena Japanese1175—
LMArena Korean1174—
LMArena Russian1234—
LMArena Spanish1228—

Instruction Following Not comparable

Gemma 2 27B: 60.5 (#249), StarCoder 2 15B: —

Instruction Following benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
LiveBench Instruction Following58.1%—
LMArena Instruction Following1206—

Long Context Not comparable

Gemma 2 27B: 37.3 (#218), StarCoder 2 15B: —

Long Context benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
LMArena Longer Query1231—

Writing & Preference Not comparable

Gemma 2 27B: 44.2 (#225), StarCoder 2 15B: —

Writing & Preference benchmarks
BenchmarkGemma 2 27BStarCoder 2 15B
LMArena Text1231—
LMArena Creative Writing1241—
LMArena Multi-Turn1224—
LiveBench Language32.6%—

Frequently asked questions

Is Gemma 2 27B better than StarCoder 2 15B?

Gemma 2 27B has enough public results to be ranked (#312); StarCoder 2 15B does not yet, so treat this comparison as directional.

Is Gemma 2 27B or StarCoder 2 15B better for coding?

StarCoder 2 15B scores higher on coding benchmarks: 36.6 versus 34.1 in the Noometry coding category.

How many benchmarks do Gemma 2 27B and StarCoder 2 15B share?

3 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and StarCoder 2 15B has 7.

Related comparisons

Go deeper