Model comparison
phi-3-medium 14B vs Phi 3 Mini 128k Instruct
phi-3-medium 14B and Phi 3 Mini 128k Instruct score almost the same on the Noometry Index (29.7 vs 29.7), so choose on price, context window or the category you care about most.
Last verified . 2 shared benchmarks.
MicrosoftSummary
- They share 2 benchmarks with published results for both. phi-3-medium 14B scores higher in 1 category and Phi 3 Mini 128k Instruct in 2 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Phi 3 Mini 128k Instruct leads 26.8 to 9.1.
- The biggest single-benchmark swing is BigCodeBench Complete: 48.7% for phi-3-medium 14B and 40.6% for Phi 3 Mini 128k Instruct.
Side by side
| phi-3-medium 14B | Phi 3 Mini 128k Instruct | |
|---|---|---|
| Provider | Microsoft | Microsoft |
| Noometry Index | 29.7 | 29.7 |
| Released | 2024-04-23 | 2024-04-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 13 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding phi-3-medium 14B leads
phi-3-medium 14B: 36.8 (#201), Phi 3 Mini 128k Instruct: 28.8 (#312)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| BigCodeBench Instruct | 37.6% | 29.6% |
| BigCodeBench Complete | 48.7% | 40.6% |
| LMArena Coding | — | 1039 |
Reasoning Not comparable
phi-3-medium 14B: —, Phi 3 Mini 128k Instruct: 19.6 (#256)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Hard Prompts | — | 1028 |
| Adversarial NLI | 55.8% | — |
| BIG-Bench Hard | 81.4% | — |
| Epoch Capabilities Index | 121.23 | — |
| HellaSwag | 82.4% | — |
| WinoGrande | 81.5% | — |
Math Phi 3 Mini 128k Instruct leads
phi-3-medium 14B: 27.3 (#250), Phi 3 Mini 128k Instruct: 31.6 (#222)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Math | — | 1089 |
| MATH Level 5 | 17.6% | — |
Knowledge Phi 3 Mini 128k Instruct leads
phi-3-medium 14B: 9.1 (#306), Phi 3 Mini 128k Instruct: 26.8 (#254)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| GPQA Diamond | 27.6% | — |
| LMArena Expert | — | 984 |
| ARC (AI2) Challenge | 91.6% | — |
| MMLU | 78% | — |
| OpenBookQA | 87.4% | — |
| TriviaQA | 73.9% | — |
Multilingual Not comparable
phi-3-medium 14B: —, Phi 3 Mini 128k Instruct: 25.2 (#285)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Non-English | — | 1000 |
| LMArena Chinese | — | 1016 |
| LMArena French | — | 1039 |
| LMArena German | — | 1006 |
| LMArena Japanese | — | 899 |
| LMArena Korean | — | 856 |
| LMArena Russian | — | 1004 |
| LMArena Spanish | — | 1059 |
Instruction Following Not comparable
phi-3-medium 14B: —, Phi 3 Mini 128k Instruct: 51.8 (#294)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Instruction Following | — | 1023 |
Long Context Not comparable
phi-3-medium 14B: —, Phi 3 Mini 128k Instruct: 30.4 (#289)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Longer Query | — | 996 |
Writing & Preference Not comparable
phi-3-medium 14B: —, Phi 3 Mini 128k Instruct: 27.1 (#301)
| Benchmark | phi-3-medium 14B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Text | — | 1050 |
| LMArena Creative Writing | — | 1024 |
| LMArena Multi-Turn | — | 989 |
Frequently asked questions
Is phi-3-medium 14B better than Phi 3 Mini 128k Instruct?
phi-3-medium 14B and Phi 3 Mini 128k Instruct score almost the same on the Noometry Index (29.7 vs 29.7), so choose on price, context window or the category you care about most.
Is phi-3-medium 14B or Phi 3 Mini 128k Instruct better for coding?
phi-3-medium 14B scores higher on coding benchmarks: 36.8 versus 28.8 in the Noometry coding category.
How many benchmarks do phi-3-medium 14B and Phi 3 Mini 128k Instruct share?
2 benchmarks have published results for both models. phi-3-medium 14B has 13 scored results on Noometry and Phi 3 Mini 128k Instruct has 19.