Model comparison
Deepseek Coder v2 vs Llama 13b
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 24.4 on the Noometry Index.
Last verified . 11 shared benchmarks.
DeepSeek
MetaSummary
- They share 11 benchmarks with published results for both. Deepseek Coder v2 scores higher in 6 categories and Llama 13b in 0 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Deepseek Coder v2 leads 61.7 to 36.7.
Side by side
| Deepseek Coder v2 | Llama 13b | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 35.9 | 24.4 |
| Released | 2024-06-17 | 2023-02-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 21 |
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), Llama 13b: 21.4 (#337)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Coding | 1251 | 683 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Llama 13b: 14.0 (#329)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Hard Prompts | 1207 | 728 |
| WinoGrande | 83.7% | 73% |
| BIG-Bench Hard | — | 37.9% |
| Epoch Capabilities Index | — | 100.58 |
| HellaSwag | — | 79.2% |
| LAMBADA | — | 75.2% |
| PIQA | — | 80.1% |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Llama 13b: 26.7 (#256)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Math | 1241 | 838 |
| GSM8K | 94.5% | 20.6% |
Knowledge Not comparable
Deepseek Coder v2: 32.3 (#212), Llama 13b: —
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| ARC (AI2) Challenge | 64.3% | 52.7% |
| LMArena Expert | 1181 | — |
| BoolQ | — | 78.7% |
| MMLU | — | 47.7% |
| OpenBookQA | — | 56.4% |
| TriviaQA | — | 77.9% |
Multimodal Not comparable
Deepseek Coder v2: —, Llama 13b: —
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual Deepseek Coder v2 leads
Deepseek Coder v2: 36.3 (#240), Llama 13b: 16.6 (#297)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Non-English | 1182 | 819 |
| LMArena Chinese | 1201 | — |
| LMArena French | 1185 | — |
| LMArena German | 1164 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Russian | 1188 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.7 (#242), Llama 13b: 36.7 (#305)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Instruction Following | 1180 | 781 |
Long Context Not comparable
Deepseek Coder v2: 37.0 (#224), Llama 13b: —
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Longer Query | 1219 | — |
Writing & Preference Deepseek Coder v2 leads
Deepseek Coder v2: 38.2 (#253), Llama 13b: 13.8 (#312)
| Benchmark | Deepseek Coder v2 | Llama 13b |
|---|---|---|
| LMArena Text | 1191 | 834 |
| LMArena Creative Writing | 1120 | 794 |
| LMArena Multi-Turn | 1177 | 753 |
Frequently asked questions
Is Deepseek Coder v2 better than Llama 13b?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 24.4 on the Noometry Index.
Is Deepseek Coder v2 or Llama 13b better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 21.4 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Llama 13b share?
11 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Llama 13b has 21.