Model comparison
Deepseek Coder v2 vs Mixtral 8x7B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 27.1 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 22 benchmarks with published results for both. Deepseek Coder v2 scores higher in 8 categories and Mixtral 8x7B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Deepseek Coder v2 leads 32.3 to 11.0.
Side by side
| Deepseek Coder v2 | Mixtral 8x7B | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 35.9 | 27.1 |
| Released | 2024-06-17 | 2023-12-11 |
| Weights | Open | Open |
| Context window | — | 32K |
| Max output | — | 32K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 24 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 38.1 (#183), Mixtral 8x7B: 32.8 (#269)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Coding | 1251 | 1126 |
| HumanEval+ | 82.3% | 39.6% |
| MBPP+ | 75.1% | 49.7% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Mixtral 8x7B: 18.2 (#285)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1115 |
| WinoGrande | 83.7% | 77.2% |
| DTBench | — | 49.6% |
| Adversarial NLI | — | 55.2% |
| Epoch Capabilities Index | — | 118.47 |
| ForecastBench | — | 56.3 |
| HellaSwag | — | 86.7% |
| PIQA | — | 83.6% |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Mixtral 8x7B: 18.8 (#289)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Math | 1241 | 1147 |
| GSM8K | 94.5% | 74.4% |
| Omni-MATH | — | 10.5% |
| MATH Level 5 | — | 10% |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Mixtral 8x7B: 11.0 (#301)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Expert | 1181 | 1088 |
| ARC (AI2) Challenge | 64.3% | 87.3% |
| GPQA Diamond | — | 30.6% |
| MMLU-Pro | — | 33.5% |
| GPQA (HELM) | — | 29.6% |
| MMLU | — | 70.6% |
| OpenBookQA | — | 85.8% |
| TriviaQA | — | 82.2% |
Multilingual Deepseek Coder v2 leads
Deepseek Coder v2: 36.3 (#240), Mixtral 8x7B: 29.6 (#266)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Non-English | 1182 | 1077 |
| LMArena Chinese | 1201 | 1055 |
| LMArena French | 1185 | 1166 |
| LMArena German | 1164 | 1114 |
| LMArena Japanese | 1126 | 931 |
| LMArena Korean | 1104 | 968 |
| LMArena Russian | 1188 | 1090 |
| LMArena Spanish | 1153 | 1111 |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.7 (#242), Mixtral 8x7B: 51.0 (#297)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1109 |
| IFEval | — | 57.5% |
Long Context Deepseek Coder v2 leads
Deepseek Coder v2: 37.0 (#224), Mixtral 8x7B: 33.4 (#260)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Longer Query | 1219 | 1103 |
Writing & Preference Deepseek Coder v2 leads
Deepseek Coder v2: 38.2 (#253), Mixtral 8x7B: 34.2 (#270)
| Benchmark | Deepseek Coder v2 | Mixtral 8x7B |
|---|---|---|
| LMArena Text | 1191 | 1132 |
| LMArena Creative Writing | 1120 | 1109 |
| LMArena Multi-Turn | 1177 | 1115 |
| WildBench | — | 67.3% |
Frequently asked questions
Is Deepseek Coder v2 better than Mixtral 8x7B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 27.1 on the Noometry Index.
Is Deepseek Coder v2 or Mixtral 8x7B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 32.8 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Mixtral 8x7B share?
22 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Mixtral 8x7B has 38.