Model comparison
Llama 2-13B vs Phi-2
Llama 2-13B has enough public results to be ranked (#304); Phi-2 does not yet, so treat this comparison as directional.
Last verified . 8 shared benchmarks.
Meta
MicrosoftSummary
- They share 8 benchmarks with published results for both.
Side by side
| Llama 2-13B | Phi-2 | |
|---|---|---|
| Provider | Meta | Microsoft |
| Noometry Index | 29.1 | — |
| Released | 2023-07-18 | 2023-12-12 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 32 | 11 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Llama 2-13B: 30.4 (#287), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Coding | 1062 | — |
| HumanEval+ | — | 45.1% |
| MBPP+ | — | 54.2% |
Reasoning Not comparable
Llama 2-13B: 12.3 (#333), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| BIG-Bench Hard | 58.2% | 59.4% |
| Epoch Capabilities Index | 105.8 | 107.62 |
| HellaSwag | 80.7% | 53.6% |
| WinoGrande | 72.8% | 54.7% |
| Chess Puzzles | 0% | — |
| LMArena Hard Prompts | 1051 | — |
| DTBench | 42.2% | — |
| Adversarial NLI | — | 42.5% |
| LAMBADA | 76.5% | — |
| PIQA | 80.8% | — |
Math Not comparable
Llama 2-13B: 28.7 (#226), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Math | 1065 | — |
| GSM8K | 36.9% | — |
Knowledge Not comparable
Llama 2-13B: 27.5 (#244), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| ARC (AI2) Challenge | 60.3% | 75.9% |
| MMLU | 55.6% | 58.4% |
| OpenBookQA | 57% | 73.6% |
| TriviaQA | 79.6% | 45.2% |
| LMArena Expert | 1030 | — |
| BoolQ | 82.4% | — |
Multimodal Not comparable
Llama 2-13B: —, Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| ScienceQA | 55.8% | — |
Multilingual Not comparable
Llama 2-13B: 26.7 (#274), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Non-English | 1024 | — |
| LMArena Chinese | 1001 | — |
| LMArena French | 1043 | — |
| LMArena German | 1009 | — |
| LMArena Japanese | 894 | — |
| LMArena Korean | 953 | — |
| LMArena Russian | 1055 | — |
| LMArena Spanish | 1086 | — |
Instruction Following Not comparable
Llama 2-13B: 53.5 (#282), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Instruction Following | 1046 | — |
Long Context Not comparable
Llama 2-13B: 32.3 (#264), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Longer Query | 1064 | — |
Writing & Preference Not comparable
Llama 2-13B: 30.0 (#284), Phi-2: —
| Benchmark | Llama 2-13B | Phi-2 |
|---|---|---|
| LMArena Text | 1085 | — |
| LMArena Creative Writing | 1047 | — |
| LMArena Multi-Turn | 1050 | — |
Frequently asked questions
Is Llama 2-13B better than Phi-2?
Llama 2-13B has enough public results to be ranked (#304); Phi-2 does not yet, so treat this comparison as directional.
How many benchmarks do Llama 2-13B and Phi-2 share?
8 benchmarks have published results for both models. Llama 2-13B has 32 scored results on Noometry and Phi-2 has 11.