Model comparison
Deepseek Coder v2 vs Llama 2-7B
Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 28.6 on the Noometry Index.
Last verified . 18 shared benchmarks.
DeepSeek
MetaSummary
- They share 18 benchmarks with published results for both. Deepseek Coder v2 scores higher in 8 categories and Llama 2-7B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Deepseek Coder v2 leads 36.4 to 23.9.
Side by side
| Deepseek Coder v2 | Llama 2-7B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 35.3 | 28.6 |
| Released | 2024-06-17 | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 29 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 37.5 (#178), Llama 2-7B: 28.7 (#302)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Coding | 1251 | 1002 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 22.6 (#175), Llama 2-7B: 15.1 (#309)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1009 |
| WinoGrande | 83.7% | 69.2% |
| Chess Puzzles | — | 0% |
| BIG-Bench Hard | — | 39.2% |
| Epoch Capabilities Index | — | 98.54 |
| HellaSwag | — | 77.2% |
| LAMBADA | — | 73.3% |
| PIQA | — | 78.8% |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#188), Llama 2-7B: 28.2 (#231)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Math | 1241 | 1042 |
| GSM8K | 94.5% | 16.7% |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 31.7 (#209), Llama 2-7B: 27.7 (#243)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Expert | 1181 | 1036 |
| ARC (AI2) Challenge | 64.3% | 45.9% |
| BoolQ | — | 77.9% |
| MMLU | — | 45.8% |
| OpenBookQA | — | 58.6% |
| TriviaQA | — | 73.7% |
Multimodal Not comparable
Deepseek Coder v2: —, Llama 2-7B: —
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| ScienceQA | — | 43.1% |
Multilingual Deepseek Coder v2 leads
Deepseek Coder v2: 36.4 (#235), Llama 2-7B: 23.9 (#288)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Non-English | 1182 | 973 |
| LMArena Chinese | 1201 | 973 |
| LMArena French | 1184 | 969 |
| LMArena German | 1164 | 978 |
| LMArena Russian | 1188 | 995 |
| LMArena Spanish | 1153 | 1007 |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.9 (#237), Llama 2-7B: 51.0 (#293)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1006 |
Long Context Deepseek Coder v2 leads
Deepseek Coder v2: 37.0 (#219), Llama 2-7B: 30.4 (#282)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Longer Query | 1219 | 999 |
Writing & Preference Deepseek Coder v2 leads
Deepseek Coder v2: 38.4 (#249), Llama 2-7B: 28.1 (#293)
| Benchmark | Deepseek Coder v2 | Llama 2-7B |
|---|---|---|
| LMArena Text | 1191 | 1053 |
| LMArena Creative Writing | 1120 | 1033 |
| LMArena Multi-Turn | 1176 | 1029 |
Frequently asked questions
Is Deepseek Coder v2 better than Llama 2-7B?
Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 28.6 on the Noometry Index.
Is Deepseek Coder v2 or Llama 2-7B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 37.5 versus 28.7 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Llama 2-7B share?
18 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Llama 2-7B has 29.