Model comparison
DBRX vs Deepseek Coder v2
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index.
Last verified . 19 shared benchmarks.
Databricks
DeepSeekSummary
- They share 19 benchmarks with published results for both. DBRX scores higher in 0 categories and Deepseek Coder v2 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Deepseek Coder v2 leads 32.3 to 14.9.
Side by side
| DBRX | Deepseek Coder v2 | |
|---|---|---|
| Provider | Databricks | DeepSeek |
| Noometry Index | 29.4 | 35.9 |
| Released | 2024-03-27 | 2024-06-17 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
DBRX: 32.9 (#266), Deepseek Coder v2: 38.1 (#183)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Coding | 1132 | 1251 |
| HumanEval+ | 70.1% | 82.3% |
| MBPP+ | 55.8% | 75.1% |
| BigCodeBench Instruct | — | 48.2% |
| BigCodeBench Complete | — | 59.7% |
Reasoning Deepseek Coder v2 leads
DBRX: 21.4 (#222), Deepseek Coder v2: 23.6 (#176)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Hard Prompts | 1113 | 1207 |
| WinoGrande | — | 83.7% |
Math Deepseek Coder v2 leads
DBRX: 24.3 (#269), Deepseek Coder v2: 34.9 (#190)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Math | 1145 | 1241 |
| MATH Level 5 | 11.7% | — |
| GSM8K | — | 94.5% |
Knowledge Deepseek Coder v2 leads
DBRX: 14.9 (#294), Deepseek Coder v2: 32.3 (#212)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Expert | 1076 | 1181 |
| GPQA Diamond | 32.9% | — |
| ARC (AI2) Challenge | — | 64.3% |
Multilingual Deepseek Coder v2 leads
DBRX: 29.3 (#268), Deepseek Coder v2: 36.3 (#240)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Non-English | 1071 | 1182 |
| LMArena Chinese | 1068 | 1201 |
| LMArena French | 1096 | 1185 |
| LMArena German | 1057 | 1164 |
| LMArena Japanese | 990 | 1126 |
| LMArena Korean | 993 | 1104 |
| LMArena Russian | 1078 | 1188 |
| LMArena Spanish | 1064 | 1153 |
Instruction Following Deepseek Coder v2 leads
DBRX: 57.5 (#270), Deepseek Coder v2: 61.7 (#242)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Instruction Following | 1112 | 1180 |
Long Context Deepseek Coder v2 leads
DBRX: 33.7 (#258), Deepseek Coder v2: 37.0 (#224)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Longer Query | 1112 | 1219 |
Writing & Preference Deepseek Coder v2 leads
DBRX: 33.6 (#275), Deepseek Coder v2: 38.2 (#253)
| Benchmark | DBRX | Deepseek Coder v2 |
|---|---|---|
| LMArena Text | 1119 | 1191 |
| LMArena Creative Writing | 1104 | 1120 |
| LMArena Multi-Turn | 1111 | 1177 |
Frequently asked questions
Is DBRX better than Deepseek Coder v2?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index.
Is DBRX or Deepseek Coder v2 better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 32.9 in the Noometry coding category.
How many benchmarks do DBRX and Deepseek Coder v2 share?
19 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and Deepseek Coder v2 has 24.