Model comparison
Phi 3 Mini 4k Instruct vs Qwen-14B
Qwen-14B is the stronger model overall, scoring 30.8 to 27.4 on the Noometry Index.
Last verified . 13 shared benchmarks.
Microsoft
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. Phi 3 Mini 4k Instruct scores higher in 1 category and Qwen-14B in 6 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen-14B leads 18.7 to 13.6.
Side by side
| Phi 3 Mini 4k Instruct | Qwen-14B | |
|---|---|---|
| Provider | Microsoft | Alibaba (Qwen) |
| Noometry Index | 27.4 | 30.8 |
| Released | 2024-04-23 | 2023-09-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen-14B leads
Phi 3 Mini 4k Instruct: 26.1 (#319), Qwen-14B: 30.6 (#283)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Coding | 1093 | 1072 |
| LiveBench Coding | 15.5% | — |
| HumanEval+ | 59.1% | — |
| MBPP+ | 54.2% | — |
Reasoning Qwen-14B leads
Phi 3 Mini 4k Instruct: 13.6 (#326), Qwen-14B: 18.7 (#255)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Hard Prompts | 1072 | 1027 |
| BIG-Bench Hard | 71.7% | 55% |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 26.8% | — |
| LiveBench Data Analysis | 34.7% | — |
| Adversarial NLI | 52.8% | — |
| Epoch Capabilities Index | — | 112.8 |
| HellaSwag | 76.7% | — |
| LAMBADA | — | 71.1% |
| LiveBench | 22.4% | — |
| PIQA | — | 79.9% |
| WinoGrande | 70.8% | — |
Math Qwen-14B leads
Phi 3 Mini 4k Instruct: 24.5 (#258), Qwen-14B: 28.7 (#224)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Math | 1111 | 1068 |
| LiveBench Math | 15.7% | — |
| GSM8K | — | 61.3% |
Knowledge Not comparable
Phi 3 Mini 4k Instruct: 27.9 (#241), Qwen-14B: —
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| ARC (AI2) Challenge | 84.9% | 84.4% |
| MMLU | 68.8% | 66.3% |
| LMArena Expert | 1045 | — |
| BoolQ | — | 86.2% |
| OpenBookQA | 88% | — |
| TriviaQA | 64% | — |
Multilingual Qwen-14B leads
Phi 3 Mini 4k Instruct: 26.5 (#275), Qwen-14B: 27.7 (#270)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Non-English | 1021 | 1042 |
| LMArena Chinese | 1021 | 1077 |
| LMArena French | 1075 | — |
| LMArena German | 1044 | — |
| LMArena Japanese | 935 | — |
| LMArena Korean | 905 | — |
| LMArena Russian | 1022 | — |
| LMArena Spanish | 1084 | — |
Instruction Following Qwen-14B leads
Phi 3 Mini 4k Instruct: 48.0 (#298), Qwen-14B: 52.6 (#284)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Instruction Following | 1053 | 1031 |
| LiveBench Instruction Following | 39.1% | — |
Long Context Too close to call
Phi 3 Mini 4k Instruct: 31.7 (#271), Qwen-14B: 31.3 (#276)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Longer Query | 1044 | 1029 |
Writing & Preference Too close to call
Phi 3 Mini 4k Instruct: 27.7 (#295), Qwen-14B: 27.8 (#294)
| Benchmark | Phi 3 Mini 4k Instruct | Qwen-14B |
|---|---|---|
| LMArena Text | 1073 | 1051 |
| LMArena Creative Writing | 1037 | 1028 |
| LMArena Multi-Turn | 1018 | 1022 |
| LiveBench Language | 9.2% | — |
Frequently asked questions
Is Phi 3 Mini 4k Instruct better than Qwen-14B?
Qwen-14B is the stronger model overall, scoring 30.8 to 27.4 on the Noometry Index.
Is Phi 3 Mini 4k Instruct or Qwen-14B better for coding?
Qwen-14B scores higher on coding benchmarks: 30.6 versus 26.1 in the Noometry coding category.
How many benchmarks do Phi 3 Mini 4k Instruct and Qwen-14B share?
13 benchmarks have published results for both models. Phi 3 Mini 4k Instruct has 35 scored results on Noometry and Qwen-14B has 18.