Model comparison
DBRX vs DeepSeek-V2 (MoE-236B, May 2024)
DBRX has enough public results to be ranked (#311); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Last verified . 0 shared benchmarks.
Databricks
DeepSeekSummary
- The widest gap is in coding, where DeepSeek-V2 (MoE-236B, May 2024) leads 40.4 to 32.9.
Side by side
| DBRX | DeepSeek-V2 (MoE-236B, May 2024) | |
|---|---|---|
| Provider | Databricks | DeepSeek |
| Noometry Index | 29.4 | 40.3 |
| Released | 2024-03-27 | 2024-05-07 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 10 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V2 (MoE-236B, May 2024) leads
DBRX: 32.9 (#266), DeepSeek-V2 (MoE-236B, May 2024): 40.4 (#139)
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| BigCodeBench Instruct | — | 48.9% |
| LMArena Coding | 1132 | — |
| BigCodeBench Complete | — | 59.4% |
| HumanEval+ | 70.1% | — |
| MBPP+ | 55.8% | — |
Reasoning Not comparable
DBRX: 21.4 (#222), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Hard Prompts | 1113 | — |
| BIG-Bench Hard | — | 78.8% |
| Epoch Capabilities Index | — | 124.77 |
| HellaSwag | — | 87.1% |
| PIQA | — | 83.9% |
| WinoGrande | — | 86.3% |
Math Not comparable
DBRX: 24.3 (#269), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Math | 1145 | — |
| MATH Level 5 | 11.7% | — |
Knowledge Not comparable
DBRX: 14.9 (#294), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| GPQA Diamond | 32.9% | — |
| LMArena Expert | 1076 | — |
| ARC (AI2) Challenge | — | 92.2% |
| MMLU | — | 78.4% |
| TriviaQA | — | 80% |
Multilingual Not comparable
DBRX: 29.3 (#268), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Non-English | 1071 | — |
| LMArena Chinese | 1068 | — |
| LMArena French | 1096 | — |
| LMArena German | 1057 | — |
| LMArena Japanese | 990 | — |
| LMArena Korean | 993 | — |
| LMArena Russian | 1078 | — |
| LMArena Spanish | 1064 | — |
Instruction Following Not comparable
DBRX: 57.5 (#270), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Instruction Following | 1112 | — |
Long Context Not comparable
DBRX: 33.7 (#258), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Longer Query | 1112 | — |
Writing & Preference Not comparable
DBRX: 33.6 (#275), DeepSeek-V2 (MoE-236B, May 2024): —
| Benchmark | DBRX | DeepSeek-V2 (MoE-236B, May 2024) |
|---|---|---|
| LMArena Text | 1119 | — |
| LMArena Creative Writing | 1104 | — |
| LMArena Multi-Turn | 1111 | — |
Frequently asked questions
Is DBRX better than DeepSeek-V2 (MoE-236B, May 2024)?
DBRX has enough public results to be ranked (#311); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Is DBRX or DeepSeek-V2 (MoE-236B, May 2024) better for coding?
DeepSeek-V2 (MoE-236B, May 2024) scores higher on coding benchmarks: 40.4 versus 32.9 in the Noometry coding category.
How many benchmarks do DBRX and DeepSeek-V2 (MoE-236B, May 2024) share?
0 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and DeepSeek-V2 (MoE-236B, May 2024) has 10.