Model comparison

Mistral vs StarCoder 2 3B

Mistral has enough public results to be ranked (#303); StarCoder 2 3B does not yet, so treat this comparison as directional.

Last verified . 0 shared benchmarks.

Mistral Mistral AI

29.9

Rank #303 Confirmed

StarCoder 2 3B NVIDIA

36.3

Unranked Sparse

Summary

  • StarCoder 2 3B has downloadable open weights; the other is API-only.

Side by side

Mistral and StarCoder 2 3B specifications
MistralStarCoder 2 3B
ProviderMistral AINVIDIA
Noometry Index29.936.3
Released—2024-02-22
WeightsProprietaryOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked227

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

Category by category

Coding Too close to call

Mistral: 33.8 (#250), StarCoder 2 3B: 33.5

Coding benchmarks
BenchmarkMistralStarCoder 2 3B
LMArena Coding1162—
BigCodeBench Complete—21.4%
HumanEval+—27.4%

Reasoning Not comparable

Mistral: 22.2 (#200), StarCoder 2 3B: —

Reasoning benchmarks
BenchmarkMistralStarCoder 2 3B
LMArena Hard Prompts1149—
Epoch Capabilities Index—88.76
WinoGrande—57.1%

Math Not comparable

Mistral: 22.3 (#278), StarCoder 2 3B: —

Math benchmarks
BenchmarkMistralStarCoder 2 3B
Omni-MATH7.2%—
LMArena Math1180—
GSM8K—21.6%

Knowledge Not comparable

Mistral: 16.6 (#288), StarCoder 2 3B: —

Knowledge benchmarks
BenchmarkMistralStarCoder 2 3B
MMLU-Pro27.7%—
GPQA (HELM)30.3%—
LMArena Expert1125—
ARC (AI2) Challenge—34.2%
MMLU—36.6%

Multilingual Not comparable

Mistral: 32.8 (#254), StarCoder 2 3B: —

Multilingual benchmarks
BenchmarkMistralStarCoder 2 3B
LMArena Non-English1129—
LMArena Chinese1109—
LMArena French1180—
LMArena German1155—
LMArena Japanese1013—
LMArena Korean1032—
LMArena Russian1168—
LMArena Spanish1143—

Instruction Following Not comparable

Mistral: 52.6 (#288), StarCoder 2 3B: —

Instruction Following benchmarks
BenchmarkMistralStarCoder 2 3B
IFEval56.8%—
LMArena Instruction Following1152—

Long Context Not comparable

Mistral: 35.0 (#245), StarCoder 2 3B: —

Long Context benchmarks
BenchmarkMistralStarCoder 2 3B
LMArena Longer Query1153—

Writing & Preference Not comparable

Mistral: 37.0 (#260), StarCoder 2 3B: —

Writing & Preference benchmarks
BenchmarkMistralStarCoder 2 3B
LMArena Text1165—
LMArena Creative Writing1158—
WildBench66%—
LMArena Multi-Turn1147—

Frequently asked questions

Is Mistral better than StarCoder 2 3B?

Mistral has enough public results to be ranked (#303); StarCoder 2 3B does not yet, so treat this comparison as directional.

Is Mistral or StarCoder 2 3B better for coding?

They score almost the same on coding (33.8 vs 33.5); test both on your own repository before choosing.

How many benchmarks do Mistral and StarCoder 2 3B share?

0 benchmarks have published results for both models. Mistral has 22 scored results on Noometry and StarCoder 2 3B has 7.

Related comparisons

Go deeper