Model comparison
Phi 3 Mini 4k Instruct vs StarCoder 2 15B
Phi 3 Mini 4k Instruct has enough public results to be ranked (#328); StarCoder 2 15B does not yet, so treat this comparison as directional.
Last verified . 4 shared benchmarks.
Microsoft
NVIDIASummary
- They share 4 benchmarks with published results for both. Phi 3 Mini 4k 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 26.6.
Side by side
| Phi 3 Mini 4k Instruct | StarCoder 2 15B | |
|---|---|---|
| Provider | Microsoft | NVIDIA |
| Noometry Index | 27.9 | 38.1 |
| Released | 2024-04-23 | 2024-02-20 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 7 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding StarCoder 2 15B leads
Phi 3 Mini 4k Instruct: 26.6 (#323), StarCoder 2 15B: 36.6
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| HumanEval+ | 59.1% | 37.8% |
| LiveBench Coding | 15.5% | — |
| LMArena Coding | 1093 | — |
| BigCodeBench Complete | — | 38.4% |
| MBPP+ | 54.2% | — |
Reasoning Not comparable
Phi 3 Mini 4k Instruct: 14.1 (#328), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| WinoGrande | 70.8% | 64.3% |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 26.8% | — |
| LMArena Hard Prompts | 1072 | — |
| LiveBench Data Analysis | 34.7% | — |
| Adversarial NLI | 52.8% | — |
| BIG-Bench Hard | 71.7% | — |
| Epoch Capabilities Index | — | 105 |
| HellaSwag | 76.7% | — |
| LiveBench | 22.4% | — |
Math Not comparable
Phi 3 Mini 4k Instruct: 26.6 (#257), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| LiveBench Math | 15.7% | — |
| LMArena Math | 1111 | — |
| GSM8K | — | 57.7% |
Knowledge Not comparable
Phi 3 Mini 4k Instruct: 28.5 (#246), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| ARC (AI2) Challenge | 84.9% | 47.2% |
| MMLU | 68.8% | 64.1% |
| LMArena Expert | 1045 | — |
| OpenBookQA | 88% | — |
| TriviaQA | 64% | — |
Multilingual Not comparable
Phi 3 Mini 4k Instruct: 26.3 (#280), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| LMArena Non-English | 1021 | — |
| LMArena Chinese | 1021 | — |
| LMArena French | 1076 | — |
| LMArena German | 1044 | — |
| LMArena Japanese | 935 | — |
| LMArena Korean | 905 | — |
| LMArena Russian | 1022 | — |
| LMArena Spanish | 1085 | — |
Instruction Following Not comparable
Phi 3 Mini 4k Instruct: 47.7 (#303), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| LiveBench Instruction Following | 39.1% | — |
| LMArena Instruction Following | 1053 | — |
Long Context Not comparable
Phi 3 Mini 4k Instruct: 31.7 (#276), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| LMArena Longer Query | 1044 | — |
Writing & Preference Not comparable
Phi 3 Mini 4k Instruct: 27.6 (#300), StarCoder 2 15B: —
| Benchmark | Phi 3 Mini 4k Instruct | StarCoder 2 15B |
|---|---|---|
| LMArena Text | 1073 | — |
| LMArena Creative Writing | 1037 | — |
| LMArena Multi-Turn | 1018 | — |
| LiveBench Language | 9.2% | — |
Frequently asked questions
Is Phi 3 Mini 4k Instruct better than StarCoder 2 15B?
Phi 3 Mini 4k Instruct has enough public results to be ranked (#328); StarCoder 2 15B does not yet, so treat this comparison as directional.
Is Phi 3 Mini 4k Instruct or StarCoder 2 15B better for coding?
StarCoder 2 15B scores higher on coding benchmarks: 36.6 versus 26.6 in the Noometry coding category.
How many benchmarks do Phi 3 Mini 4k Instruct and StarCoder 2 15B share?
4 benchmarks have published results for both models. Phi 3 Mini 4k Instruct has 35 scored results on Noometry and StarCoder 2 15B has 7.