Model comparison
DBRX vs Llama 2-70B
DBRX is the stronger model overall, scoring 29.4 to 24.4 on the Noometry Index.
Last verified . 19 shared benchmarks.
Databricks
MetaSummary
- They share 19 benchmarks with published results for both. DBRX scores higher in 8 categories and Llama 2-70B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DBRX leads 24.3 to 8.1.
- The biggest single-benchmark swing is MATH Level 5: 11.7% for DBRX and 3.3% for Llama 2-70B.
Side by side
| DBRX | Llama 2-70B | |
|---|---|---|
| Provider | Databricks | Meta |
| Noometry Index | 29.4 | 24.4 |
| Released | 2024-03-27 | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DBRX leads
DBRX: 32.9 (#266), Llama 2-70B: 31.4 (#286)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1132 | 1079 |
| HumanEval+ | 70.1% | — |
| MBPP+ | 55.8% | — |
Reasoning DBRX leads
DBRX: 21.4 (#222), Llama 2-70B: 14.4 (#325)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1113 | 1073 |
| DTBench | — | 41.6% |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| Epoch Capabilities Index | — | 113.79 |
| ForecastBench | — | 51.4 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |
Math DBRX leads
DBRX: 24.3 (#269), Llama 2-70B: 8.1 (#326)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Math | 1145 | 1091 |
| MATH Level 5 | 11.7% | 3.3% |
| OTIS Mock AIME 2024-2025 | — | 0% |
| GSM8K | — | 69.6% |
Knowledge DBRX leads
DBRX: 14.9 (#294), Llama 2-70B: 7.4 (#310)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| GPQA Diamond | 32.9% | 26.3% |
| LMArena Expert | 1076 | 1039 |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multilingual DBRX leads
DBRX: 29.3 (#268), Llama 2-70B: 27.7 (#274)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1071 | 1045 |
| LMArena Chinese | 1068 | 995 |
| LMArena French | 1096 | 1090 |
| LMArena German | 1057 | 1041 |
| LMArena Japanese | 990 | 927 |
| LMArena Korean | 993 | 964 |
| LMArena Russian | 1078 | 1083 |
| LMArena Spanish | 1064 | 1143 |
Instruction Following DBRX leads
DBRX: 57.5 (#270), Llama 2-70B: 54.9 (#278)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1112 | 1071 |
Long Context DBRX leads
DBRX: 33.7 (#258), Llama 2-70B: 32.3 (#270)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1112 | 1062 |
Writing & Preference DBRX leads
DBRX: 33.6 (#275), Llama 2-70B: 32.3 (#279)
| Benchmark | DBRX | Llama 2-70B |
|---|---|---|
| LMArena Text | 1119 | 1115 |
| LMArena Creative Writing | 1104 | 1075 |
| LMArena Multi-Turn | 1111 | 1088 |
Frequently asked questions
Is DBRX better than Llama 2-70B?
DBRX is the stronger model overall, scoring 29.4 to 24.4 on the Noometry Index.
Is DBRX or Llama 2-70B better for coding?
DBRX scores higher on coding benchmarks: 32.9 versus 31.4 in the Noometry coding category.
How many benchmarks do DBRX and Llama 2-70B share?
19 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and Llama 2-70B has 35.