Model comparison
Llama 2-7B vs Phi 3 Mini 4k Instruct
Llama 2-7B is the stronger model overall, scoring 28.6 to 27.4 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
MicrosoftSummary
- They share 23 benchmarks with published results for both. Llama 2-7B scores higher in 5 categories and Phi 3 Mini 4k Instruct in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Llama 2-7B leads 28.2 to 24.5.
Side by side
| Llama 2-7B | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | Meta | Microsoft |
| Noometry Index | 28.6 | 27.4 |
| Released | 2023-07-18 | 2024-04-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 29 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-7B leads
Llama 2-7B: 28.7 (#302), Phi 3 Mini 4k Instruct: 26.1 (#319)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1002 | 1093 |
| LiveBench Coding | — | 15.5% |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Reasoning Llama 2-7B leads
Llama 2-7B: 15.1 (#309), Phi 3 Mini 4k Instruct: 13.6 (#326)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1009 | 1072 |
| BIG-Bench Hard | 39.2% | 71.7% |
| HellaSwag | 77.2% | 76.7% |
| WinoGrande | 69.2% | 70.8% |
| LiveBench Reasoning | — | 26.8% |
| LiveBench Data Analysis | — | 34.7% |
| Adversarial NLI | — | 52.8% |
| Epoch Capabilities Index | 98.54 | — |
| LAMBADA | 73.3% | — |
| LiveBench | — | 22.4% |
| PIQA | 78.8% | — |
Math Llama 2-7B leads
Llama 2-7B: 28.2 (#231), Phi 3 Mini 4k Instruct: 24.5 (#258)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1042 | 1111 |
| LiveBench Math | — | 15.7% |
| GSM8K | 16.7% | — |
Knowledge Too close to call
Llama 2-7B: 27.7 (#243), Phi 3 Mini 4k Instruct: 27.9 (#241)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1036 | 1045 |
| ARC (AI2) Challenge | 45.9% | 84.9% |
| MMLU | 45.8% | 68.8% |
| OpenBookQA | 58.6% | 88% |
| TriviaQA | 73.7% | 64% |
| BoolQ | 77.9% | — |
Multimodal Not comparable
Llama 2-7B: —, Phi 3 Mini 4k Instruct: —
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| ScienceQA | 43.1% | — |
Multilingual Phi 3 Mini 4k Instruct leads
Llama 2-7B: 23.9 (#288), Phi 3 Mini 4k Instruct: 26.5 (#275)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 973 | 1021 |
| LMArena Chinese | 973 | 1021 |
| LMArena French | 969 | 1075 |
| LMArena German | 978 | 1044 |
| LMArena Russian | 995 | 1022 |
| LMArena Spanish | 1007 | 1084 |
| LMArena Japanese | — | 935 |
| LMArena Korean | — | 905 |
Instruction Following Llama 2-7B leads
Llama 2-7B: 51.0 (#293), Phi 3 Mini 4k Instruct: 48.0 (#298)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1006 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context Phi 3 Mini 4k Instruct leads
Llama 2-7B: 30.4 (#282), Phi 3 Mini 4k Instruct: 31.7 (#271)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 999 | 1044 |
Writing & Preference Too close to call
Llama 2-7B: 28.1 (#293), Phi 3 Mini 4k Instruct: 27.7 (#295)
| Benchmark | Llama 2-7B | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1053 | 1073 |
| LMArena Creative Writing | 1033 | 1037 |
| LMArena Multi-Turn | 1029 | 1018 |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is Llama 2-7B better than Phi 3 Mini 4k Instruct?
Llama 2-7B is the stronger model overall, scoring 28.6 to 27.4 on the Noometry Index.
Is Llama 2-7B or Phi 3 Mini 4k Instruct better for coding?
Llama 2-7B scores higher on coding benchmarks: 28.7 versus 26.1 in the Noometry coding category.
How many benchmarks do Llama 2-7B and Phi 3 Mini 4k Instruct share?
23 benchmarks have published results for both models. Llama 2-7B has 29 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.