Model comparison
DeepSeek-V2.5 (Sep 2024) vs Llama 2-70B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 24.4 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
MetaSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 8 categories and Llama 2-70B in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 35.9 to 8.1.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Llama 2-70B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 37.6 | 24.4 |
| Released | 2024-09-06 | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Llama 2-70B: 31.4 (#286)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1309 | 1079 |
| Aider Polyglot | 17.8% | — |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Llama 2-70B: 14.4 (#325)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1073 |
| DTBench | — | 41.6% |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| Epoch Capabilities Index | — | 113.79 |
| ForecastBench | — | 51.4 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Llama 2-70B: 8.1 (#326)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Math | 1288 | 1091 |
| OTIS Mock AIME 2024-2025 | — | 0% |
| MATH Level 5 | — | 3.3% |
| GSM8K | — | 69.6% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Llama 2-70B: 7.4 (#310)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Expert | 1266 | 1039 |
| GPQA Diamond | — | 26.3% |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Llama 2-70B: 27.7 (#274)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1273 | 1045 |
| LMArena Chinese | 1318 | 995 |
| LMArena French | 1289 | 1090 |
| LMArena German | 1258 | 1041 |
| LMArena Japanese | 1228 | 927 |
| LMArena Korean | 1209 | 964 |
| LMArena Russian | 1289 | 1083 |
| LMArena Spanish | 1248 | 1143 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Llama 2-70B: 54.9 (#278)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1071 |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Llama 2-70B: 32.3 (#270)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1301 | 1062 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Llama 2-70B: 32.3 (#279)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 2-70B |
|---|---|---|
| LMArena Text | 1294 | 1115 |
| LMArena Creative Writing | 1285 | 1075 |
| LMArena Multi-Turn | 1297 | 1088 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Llama 2-70B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 24.4 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Llama 2-70B better for coding?
They score almost the same on coding (31.7 vs 31.4); test both on your own repository before choosing.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Llama 2-70B share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Llama 2-70B has 35.