Model comparison
Codellama 34b Instruct vs Deepseek Coder v2
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 30.8 on the Noometry Index.
Last verified . 14 shared benchmarks.
Meta
DeepSeekSummary
- They share 14 benchmarks with published results for both. Codellama 34b Instruct scores higher in 0 categories and Deepseek Coder v2 in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Deepseek Coder v2 leads 36.3 to 25.8.
- The biggest single-benchmark swing is BigCodeBench Complete: 37.1% for Codellama 34b Instruct and 59.7% for Deepseek Coder v2.
Side by side
| Codellama 34b Instruct | Deepseek Coder v2 | |
|---|---|---|
| Provider | Meta | DeepSeek |
| Noometry Index | 30.8 | 35.9 |
| Released | — | 2024-06-17 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 14 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Codellama 34b Instruct: 28.5 (#314), Deepseek Coder v2: 38.1 (#183)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| BigCodeBench Instruct | 29% | 48.2% |
| LMArena Coding | 1046 | 1251 |
| BigCodeBench Complete | 37.1% | 59.7% |
| HumanEval+ | 43.9% | 82.3% |
| MBPP+ | 56.3% | 75.1% |
Reasoning Deepseek Coder v2 leads
Codellama 34b Instruct: 19.6 (#255), Deepseek Coder v2: 23.6 (#176)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Hard Prompts | 1032 | 1207 |
| WinoGrande | — | 83.7% |
Math Deepseek Coder v2 leads
Codellama 34b Instruct: 31.0 (#230), Deepseek Coder v2: 34.9 (#190)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Math | 1056 | 1241 |
| GSM8K | — | 94.5% |
Knowledge Not comparable
Codellama 34b Instruct: —, Deepseek Coder v2: 32.3 (#212)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Expert | — | 1181 |
| ARC (AI2) Challenge | — | 64.3% |
Multilingual Deepseek Coder v2 leads
Codellama 34b Instruct: 25.8 (#284), Deepseek Coder v2: 36.3 (#240)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Non-English | 1011 | 1182 |
| LMArena Chinese | 976 | 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
Codellama 34b Instruct: 52.2 (#291), Deepseek Coder v2: 61.7 (#242)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Instruction Following | 1028 | 1180 |
Long Context Deepseek Coder v2 leads
Codellama 34b Instruct: 30.9 (#284), Deepseek Coder v2: 37.0 (#224)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Longer Query | 1013 | 1219 |
Writing & Preference Deepseek Coder v2 leads
Codellama 34b Instruct: 28.2 (#297), Deepseek Coder v2: 38.2 (#253)
| Benchmark | Codellama 34b Instruct | Deepseek Coder v2 |
|---|---|---|
| LMArena Text | 1066 | 1191 |
| LMArena Creative Writing | 1032 | 1120 |
| LMArena Multi-Turn | 1015 | 1177 |
Frequently asked questions
Is Codellama 34b Instruct better than Deepseek Coder v2?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 30.8 on the Noometry Index.
Is Codellama 34b Instruct or Deepseek Coder v2 better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 28.5 in the Noometry coding category.
How many benchmarks do Codellama 34b Instruct and Deepseek Coder v2 share?
14 benchmarks have published results for both models. Codellama 34b Instruct has 14 scored results on Noometry and Deepseek Coder v2 has 24.