Model comparison
Llama 13b vs phi-3-medium 14B
phi-3-medium 14B is the stronger model overall, scoring 29.7 to 24.4 on the Noometry Index.
Last verified . 8 shared benchmarks.
Meta
MicrosoftSummary
- They share 8 benchmarks with published results for both. Llama 13b scores higher in 0 categories and phi-3-medium 14B in 2 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where phi-3-medium 14B leads 36.8 to 21.4.
Side by side
| Llama 13b | phi-3-medium 14B | |
|---|---|---|
| Provider | Meta | Microsoft |
| Noometry Index | 24.4 | 29.7 |
| Released | 2023-02-24 | 2024-04-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 13 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding phi-3-medium 14B leads
Llama 13b: 21.4 (#337), phi-3-medium 14B: 36.8 (#201)
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| BigCodeBench Instruct | — | 37.6% |
| LMArena Coding | 683 | — |
| BigCodeBench Complete | — | 48.7% |
Reasoning Not comparable
Llama 13b: 14.0 (#329), phi-3-medium 14B: —
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| BIG-Bench Hard | 37.9% | 81.4% |
| Epoch Capabilities Index | 100.58 | 121.23 |
| HellaSwag | 79.2% | 82.4% |
| WinoGrande | 73% | 81.5% |
| LMArena Hard Prompts | 728 | — |
| Adversarial NLI | — | 55.8% |
| LAMBADA | 75.2% | — |
| PIQA | 80.1% | — |
Math Too close to call
Llama 13b: 26.7 (#256), phi-3-medium 14B: 27.3 (#250)
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| LMArena Math | 838 | — |
| MATH Level 5 | — | 17.6% |
| GSM8K | 20.6% | — |
Knowledge Not comparable
Llama 13b: —, phi-3-medium 14B: 9.1 (#306)
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| ARC (AI2) Challenge | 52.7% | 91.6% |
| MMLU | 47.7% | 78% |
| OpenBookQA | 56.4% | 87.4% |
| TriviaQA | 77.9% | 73.9% |
| GPQA Diamond | — | 27.6% |
| BoolQ | 78.7% | — |
Multimodal Not comparable
Llama 13b: —, phi-3-medium 14B: —
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| ScienceQA | 43.3% | — |
Multilingual Not comparable
Llama 13b: 16.6 (#297), phi-3-medium 14B: —
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| LMArena Non-English | 819 | — |
Instruction Following Not comparable
Llama 13b: 36.7 (#305), phi-3-medium 14B: —
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| LMArena Instruction Following | 781 | — |
Writing & Preference Not comparable
Llama 13b: 13.8 (#312), phi-3-medium 14B: —
| Benchmark | Llama 13b | phi-3-medium 14B |
|---|---|---|
| LMArena Text | 834 | — |
| LMArena Creative Writing | 794 | — |
| LMArena Multi-Turn | 753 | — |
Frequently asked questions
Is Llama 13b better than phi-3-medium 14B?
phi-3-medium 14B is the stronger model overall, scoring 29.7 to 24.4 on the Noometry Index.
Is Llama 13b or phi-3-medium 14B better for coding?
phi-3-medium 14B scores higher on coding benchmarks: 36.8 versus 21.4 in the Noometry coding category.
How many benchmarks do Llama 13b and phi-3-medium 14B share?
8 benchmarks have published results for both models. Llama 13b has 21 scored results on Noometry and phi-3-medium 14B has 13.