Model comparison

Codellama 34b Instruct vs StarCoder 2 15B

Codellama 34b Instruct has enough public results to be ranked (#287); StarCoder 2 15B does not yet, so treat this comparison as directional.

Last verified . 2 shared benchmarks.

Codellama 34b Instruct Meta

30.8

Rank #287 Confirmed

StarCoder 2 15B NVIDIA

38.1

Unranked Sparse

Summary

  • They share 2 benchmarks with published results for both. Codellama 34b Instruct scores higher in 0 categories and StarCoder 2 15B in 1 category; one gap is clear of the uncertainty.
  • The widest gap is in coding, where StarCoder 2 15B leads 36.6 to 28.5.

Side by side

Codellama 34b Instruct and StarCoder 2 15B specifications
Codellama 34b InstructStarCoder 2 15B
ProviderMetaNVIDIA
Noometry Index30.838.1
Released—2024-02-20
WeightsOpenOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked147

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

Category by category

Coding StarCoder 2 15B leads

Codellama 34b Instruct: 28.5 (#314), StarCoder 2 15B: 36.6

Coding benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
BigCodeBench Complete37.1%38.4%
HumanEval+43.9%37.8%
BigCodeBench Instruct29%—
LMArena Coding1046—
MBPP+56.3%—

Reasoning Not comparable

Codellama 34b Instruct: 19.6 (#255), StarCoder 2 15B: —

Reasoning benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Hard Prompts1032—
Epoch Capabilities Index—105
WinoGrande—64.3%

Math Not comparable

Codellama 34b Instruct: 31.0 (#230), StarCoder 2 15B: —

Math benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Math1056—
GSM8K—57.7%

Knowledge Not comparable

Codellama 34b Instruct: —, StarCoder 2 15B: —

Knowledge benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
ARC (AI2) Challenge—47.2%
MMLU—64.1%

Multilingual Not comparable

Codellama 34b Instruct: 25.8 (#284), StarCoder 2 15B: —

Multilingual benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Non-English1011—
LMArena Chinese976—

Instruction Following Not comparable

Codellama 34b Instruct: 52.2 (#291), StarCoder 2 15B: —

Instruction Following benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Instruction Following1028—

Long Context Not comparable

Codellama 34b Instruct: 30.9 (#284), StarCoder 2 15B: —

Long Context benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Longer Query1013—

Writing & Preference Not comparable

Codellama 34b Instruct: 28.2 (#297), StarCoder 2 15B: —

Writing & Preference benchmarks
BenchmarkCodellama 34b InstructStarCoder 2 15B
LMArena Text1066—
LMArena Creative Writing1032—
LMArena Multi-Turn1015—

Frequently asked questions

Is Codellama 34b Instruct better than StarCoder 2 15B?

Codellama 34b Instruct has enough public results to be ranked (#287); StarCoder 2 15B does not yet, so treat this comparison as directional.

Is Codellama 34b Instruct or StarCoder 2 15B better for coding?

StarCoder 2 15B scores higher on coding benchmarks: 36.6 versus 28.5 in the Noometry coding category.

How many benchmarks do Codellama 34b Instruct and StarCoder 2 15B share?

2 benchmarks have published results for both models. Codellama 34b Instruct has 14 scored results on Noometry and StarCoder 2 15B has 7.

Related comparisons

Go deeper