Model comparison
DeepSeek-V2.5 (Sep 2024) vs Mixtral 8x7B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 27.1 on the Noometry Index.
Last verified . 19 shared benchmarks.
DeepSeek
Mistral AISummary
- They share 19 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 7 categories and Mixtral 8x7B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V2.5 (Sep 2024) leads 34.8 to 11.0.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B | |
|---|---|---|
| Provider | DeepSeek | Mistral AI |
| Noometry Index | 37.6 | 27.1 |
| Released | 2024-09-06 | 2023-12-11 |
| Weights | Open | Open |
| Context window | — | 32K |
| Max output | — | 32K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 22 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mixtral 8x7B leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Mixtral 8x7B: 32.8 (#269)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Coding | 1309 | 1126 |
| HumanEval+ | 83.5% | 39.6% |
| MBPP+ | 74.1% | 49.7% |
| Aider Polyglot | 17.8% | — |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Mixtral 8x7B: 18.2 (#285)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1115 |
| 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 DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Mixtral 8x7B: 18.8 (#289)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Math | 1288 | 1147 |
| Omni-MATH | — | 10.5% |
| MATH Level 5 | — | 10% |
| GSM8K | — | 74.4% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Mixtral 8x7B: 11.0 (#301)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Expert | 1266 | 1088 |
| 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 DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Mixtral 8x7B: 29.6 (#266)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Non-English | 1273 | 1077 |
| LMArena Chinese | 1318 | 1055 |
| LMArena French | 1289 | 1166 |
| LMArena German | 1258 | 1114 |
| LMArena Japanese | 1228 | 931 |
| LMArena Korean | 1209 | 968 |
| LMArena Russian | 1289 | 1090 |
| LMArena Spanish | 1248 | 1111 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Mixtral 8x7B: 51.0 (#297)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1109 |
| IFEval | — | 57.5% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Mixtral 8x7B: 33.4 (#260)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Longer Query | 1301 | 1103 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Mixtral 8x7B: 34.2 (#270)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Mixtral 8x7B |
|---|---|---|
| LMArena Text | 1294 | 1132 |
| LMArena Creative Writing | 1285 | 1109 |
| LMArena Multi-Turn | 1297 | 1115 |
| WildBench | — | 67.3% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Mixtral 8x7B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 27.1 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Mixtral 8x7B better for coding?
Mixtral 8x7B scores higher on coding benchmarks: 32.8 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Mixtral 8x7B share?
19 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Mixtral 8x7B has 38.