Model comparison
Llama 2-13B vs Phi 3 Mini 4k Instruct
Llama 2-13B is the stronger model overall, scoring 29.6 to 27.9 on the Noometry Index.
Last verified . 25 shared benchmarks.
Meta
MicrosoftSummary
- They share 25 benchmarks with published results for both. Llama 2-13B scores higher in 6 categories and Phi 3 Mini 4k Instruct in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Llama 2-13B leads 53.3 to 47.7.
Side by side
| Llama 2-13B | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | Meta | Microsoft |
| Noometry Index | 29.6 | 27.9 |
| Released | 2023-07-18 | 2024-04-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 32 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-13B leads
Llama 2-13B: 30.9 (#291), Phi 3 Mini 4k Instruct: 26.6 (#323)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1062 | 1093 |
| LiveBench Coding | — | 15.5% |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Reasoning Phi 3 Mini 4k Instruct leads
Llama 2-13B: 12.8 (#337), Phi 3 Mini 4k Instruct: 14.1 (#328)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1051 | 1072 |
| BIG-Bench Hard | 58.2% | 71.7% |
| HellaSwag | 80.7% | 76.7% |
| WinoGrande | 72.8% | 70.8% |
| LiveBench Reasoning | — | 26.8% |
| DTBench | 42.2% | — |
| LiveBench Data Analysis | — | 34.7% |
| Adversarial NLI | — | 52.8% |
| Epoch Capabilities Index | 106.17 | — |
| LAMBADA | 76.5% | — |
| LiveBench | — | 22.4% |
| PIQA | 80.8% | — |
Math Llama 2-13B leads
Llama 2-13B: 31.1 (#229), Phi 3 Mini 4k Instruct: 26.6 (#257)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1065 | 1111 |
| LiveBench Math | — | 15.7% |
| GSM8K | 36.9% | — |
Knowledge Too close to call
Llama 2-13B: 28.1 (#249), Phi 3 Mini 4k Instruct: 28.5 (#246)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1030 | 1045 |
| ARC (AI2) Challenge | 60.3% | 84.9% |
| MMLU | 55.6% | 68.8% |
| OpenBookQA | 57% | 88% |
| TriviaQA | 79.6% | 64% |
| BoolQ | 82.4% | — |
Multimodal Not comparable
Llama 2-13B: —, Phi 3 Mini 4k Instruct: —
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| ScienceQA | 55.8% | — |
Multilingual Too close to call
Llama 2-13B: 26.5 (#279), Phi 3 Mini 4k Instruct: 26.3 (#280)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1024 | 1021 |
| LMArena Chinese | 1001 | 1021 |
| LMArena French | 1044 | 1076 |
| LMArena German | 1009 | 1044 |
| LMArena Japanese | 894 | 935 |
| LMArena Korean | 953 | 905 |
| LMArena Russian | 1055 | 1022 |
| LMArena Spanish | 1087 | 1085 |
Instruction Following Llama 2-13B leads
Llama 2-13B: 53.3 (#287), Phi 3 Mini 4k Instruct: 47.7 (#303)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1045 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context Too close to call
Llama 2-13B: 32.3 (#269), Phi 3 Mini 4k Instruct: 31.7 (#276)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1064 | 1044 |
Writing & Preference Llama 2-13B leads
Llama 2-13B: 29.8 (#289), Phi 3 Mini 4k Instruct: 27.6 (#300)
| Benchmark | Llama 2-13B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1084 | 1073 |
| LMArena Creative Writing | 1047 | 1037 |
| LMArena Multi-Turn | 1050 | 1018 |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is Llama 2-13B better than Phi 3 Mini 4k Instruct?
Llama 2-13B is the stronger model overall, scoring 29.6 to 27.9 on the Noometry Index.
Is Llama 2-13B or Phi 3 Mini 4k Instruct better for coding?
Llama 2-13B scores higher on coding benchmarks: 30.9 versus 26.6 in the Noometry coding category.
How many benchmarks do Llama 2-13B and Phi 3 Mini 4k Instruct share?
25 benchmarks have published results for both models. Llama 2-13B has 32 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.