Model comparison
DBRX vs Llama 13b
DBRX is the stronger model overall, scoring 29.4 to 24.4 on the Noometry Index.
Last verified . 8 shared benchmarks.
Databricks
MetaSummary
- They share 8 benchmarks with published results for both. DBRX scores higher in 5 categories and Llama 13b in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where DBRX leads 57.5 to 36.7.
Side by side
| DBRX | Llama 13b | |
|---|---|---|
| Provider | Databricks | Meta |
| Noometry Index | 29.4 | 24.4 |
| Released | 2024-03-27 | 2023-02-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DBRX leads
DBRX: 32.9 (#266), Llama 13b: 21.4 (#337)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Coding | 1132 | 683 |
| HumanEval+ | 70.1% | — |
| MBPP+ | 55.8% | — |
Reasoning DBRX leads
DBRX: 21.4 (#222), Llama 13b: 14.0 (#329)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Hard Prompts | 1113 | 728 |
| BIG-Bench Hard | — | 37.9% |
| Epoch Capabilities Index | — | 100.58 |
| HellaSwag | — | 79.2% |
| LAMBADA | — | 75.2% |
| PIQA | — | 80.1% |
| WinoGrande | — | 73% |
Math Llama 13b leads
DBRX: 24.3 (#269), Llama 13b: 26.7 (#256)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Math | 1145 | 838 |
| MATH Level 5 | 11.7% | — |
| GSM8K | — | 20.6% |
Knowledge Not comparable
DBRX: 14.9 (#294), Llama 13b: —
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| GPQA Diamond | 32.9% | — |
| LMArena Expert | 1076 | — |
| ARC (AI2) Challenge | — | 52.7% |
| BoolQ | — | 78.7% |
| MMLU | — | 47.7% |
| OpenBookQA | — | 56.4% |
| TriviaQA | — | 77.9% |
Multimodal Not comparable
DBRX: —, Llama 13b: —
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual DBRX leads
DBRX: 29.3 (#268), Llama 13b: 16.6 (#297)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Non-English | 1071 | 819 |
| LMArena Chinese | 1068 | — |
| LMArena French | 1096 | — |
| LMArena German | 1057 | — |
| LMArena Japanese | 990 | — |
| LMArena Korean | 993 | — |
| LMArena Russian | 1078 | — |
| LMArena Spanish | 1064 | — |
Instruction Following DBRX leads
DBRX: 57.5 (#270), Llama 13b: 36.7 (#305)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Instruction Following | 1112 | 781 |
Long Context Not comparable
DBRX: 33.7 (#258), Llama 13b: —
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Longer Query | 1112 | — |
Writing & Preference DBRX leads
DBRX: 33.6 (#275), Llama 13b: 13.8 (#312)
| Benchmark | DBRX | Llama 13b |
|---|---|---|
| LMArena Text | 1119 | 834 |
| LMArena Creative Writing | 1104 | 794 |
| LMArena Multi-Turn | 1111 | 753 |
Frequently asked questions
Is DBRX better than Llama 13b?
DBRX is the stronger model overall, scoring 29.4 to 24.4 on the Noometry Index.
Is DBRX or Llama 13b better for coding?
DBRX scores higher on coding benchmarks: 32.9 versus 21.4 in the Noometry coding category.
How many benchmarks do DBRX and Llama 13b share?
8 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and Llama 13b has 21.