Model comparison
Mixtral 8x7B vs Phi 3 Mini 128k Instruct
Phi 3 Mini 128k Instruct is the stronger model overall, scoring 29.7 to 27.1 on the Noometry Index.
Last verified . 17 shared benchmarks.
Summary
- They share 17 benchmarks with published results for both. Mixtral 8x7B scores higher in 4 categories and Phi 3 Mini 128k Instruct in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Phi 3 Mini 128k Instruct leads 26.8 to 11.0.
Side by side
| Mixtral 8x7B | Phi 3 Mini 128k Instruct | |
|---|---|---|
| Provider | Mistral AI | Microsoft |
| Noometry Index | 27.1 | 29.7 |
| Released | 2023-12-11 | 2024-04-23 |
| Weights | Open | Open |
| Context window | 32K | — |
| Max output | 32K | — |
| Input $ / M tokens | $0.70 | — |
| Output $ / M tokens | $0.70 | — |
| Results tracked | 38 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mixtral 8x7B leads
Mixtral 8x7B: 32.8 (#269), Phi 3 Mini 128k Instruct: 28.8 (#312)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Coding | 1126 | 1039 |
| BigCodeBench Instruct | — | 29.6% |
| BigCodeBench Complete | — | 40.6% |
| HumanEval+ | 39.6% | — |
| MBPP+ | 49.7% | — |
Reasoning Phi 3 Mini 128k Instruct leads
Mixtral 8x7B: 18.2 (#285), Phi 3 Mini 128k Instruct: 19.6 (#256)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Hard Prompts | 1115 | 1028 |
| DTBench | 49.6% | — |
| Adversarial NLI | 55.2% | — |
| Epoch Capabilities Index | 118.47 | — |
| ForecastBench | 56.3 | — |
| HellaSwag | 86.7% | — |
| PIQA | 83.6% | — |
| WinoGrande | 77.2% | — |
Math Phi 3 Mini 128k Instruct leads
Mixtral 8x7B: 18.8 (#289), Phi 3 Mini 128k Instruct: 31.6 (#222)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Math | 1147 | 1089 |
| Omni-MATH | 10.5% | — |
| MATH Level 5 | 10% | — |
| GSM8K | 74.4% | — |
Knowledge Phi 3 Mini 128k Instruct leads
Mixtral 8x7B: 11.0 (#301), Phi 3 Mini 128k Instruct: 26.8 (#254)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Expert | 1088 | 984 |
| GPQA Diamond | 30.6% | — |
| MMLU-Pro | 33.5% | — |
| GPQA (HELM) | 29.6% | — |
| ARC (AI2) Challenge | 87.3% | — |
| MMLU | 70.6% | — |
| OpenBookQA | 85.8% | — |
| TriviaQA | 82.2% | — |
Multilingual Mixtral 8x7B leads
Mixtral 8x7B: 29.6 (#266), Phi 3 Mini 128k Instruct: 25.2 (#285)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Non-English | 1077 | 1000 |
| LMArena Chinese | 1055 | 1016 |
| LMArena French | 1166 | 1039 |
| LMArena German | 1114 | 1006 |
| LMArena Japanese | 931 | 899 |
| LMArena Korean | 968 | 856 |
| LMArena Russian | 1090 | 1004 |
| LMArena Spanish | 1111 | 1059 |
Instruction Following Too close to call
Mixtral 8x7B: 51.0 (#297), Phi 3 Mini 128k Instruct: 51.8 (#294)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Instruction Following | 1109 | 1023 |
| IFEval | 57.5% | — |
Long Context Mixtral 8x7B leads
Mixtral 8x7B: 33.4 (#260), Phi 3 Mini 128k Instruct: 30.4 (#289)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Longer Query | 1103 | 996 |
Writing & Preference Mixtral 8x7B leads
Mixtral 8x7B: 34.2 (#270), Phi 3 Mini 128k Instruct: 27.1 (#301)
| Benchmark | Mixtral 8x7B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Text | 1132 | 1050 |
| LMArena Creative Writing | 1109 | 1024 |
| LMArena Multi-Turn | 1115 | 989 |
| WildBench | 67.3% | — |
Frequently asked questions
Is Mixtral 8x7B better than Phi 3 Mini 128k Instruct?
Phi 3 Mini 128k Instruct is the stronger model overall, scoring 29.7 to 27.1 on the Noometry Index.
Is Mixtral 8x7B or Phi 3 Mini 128k Instruct better for coding?
Mixtral 8x7B scores higher on coding benchmarks: 32.8 versus 28.8 in the Noometry coding category.
How many benchmarks do Mixtral 8x7B and Phi 3 Mini 128k Instruct share?
17 benchmarks have published results for both models. Mixtral 8x7B has 38 scored results on Noometry and Phi 3 Mini 128k Instruct has 19.