Model comparison
DBRX vs DeepSeek-V2.5 (Sep 2024)
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 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 1 category and DeepSeek-V2.5 (Sep 2024) in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V2.5 (Sep 2024) leads 34.8 to 14.9.
Side by side
| DBRX | DeepSeek-V2.5 (Sep 2024) | |
|---|---|---|
| Provider | Databricks | DeepSeek |
| Noometry Index | 29.4 | 37.6 |
| Released | 2024-03-27 | 2024-09-06 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DBRX leads
DBRX: 32.9 (#266), DeepSeek-V2.5 (Sep 2024): 31.7 (#281)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Coding | 1132 | 1309 |
| HumanEval+ | 70.1% | 83.5% |
| MBPP+ | 55.8% | 74.1% |
| Aider Polyglot | — | 17.8% |
| BigCodeBench Instruct | — | 48.6% |
| BigCodeBench Complete | — | 53.2% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DBRX: 21.4 (#222), DeepSeek-V2.5 (Sep 2024): 25.6 (#145)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Hard Prompts | 1113 | 1289 |
Math DeepSeek-V2.5 (Sep 2024) leads
DBRX: 24.3 (#269), DeepSeek-V2.5 (Sep 2024): 35.9 (#177)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Math | 1145 | 1288 |
| MATH Level 5 | 11.7% | — |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DBRX: 14.9 (#294), DeepSeek-V2.5 (Sep 2024): 34.8 (#193)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Expert | 1076 | 1266 |
| GPQA Diamond | 32.9% | — |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DBRX: 29.3 (#268), DeepSeek-V2.5 (Sep 2024): 42.5 (#193)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Non-English | 1071 | 1273 |
| LMArena Chinese | 1068 | 1318 |
| LMArena French | 1096 | 1289 |
| LMArena German | 1057 | 1258 |
| LMArena Japanese | 990 | 1228 |
| LMArena Korean | 993 | 1209 |
| LMArena Russian | 1078 | 1289 |
| LMArena Spanish | 1064 | 1248 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DBRX: 57.5 (#270), DeepSeek-V2.5 (Sep 2024): 67.5 (#194)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Instruction Following | 1112 | 1280 |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DBRX: 33.7 (#258), DeepSeek-V2.5 (Sep 2024): 39.5 (#174)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Longer Query | 1112 | 1301 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DBRX: 33.6 (#275), DeepSeek-V2.5 (Sep 2024): 49.8 (#187)
| Benchmark | DBRX | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Text | 1119 | 1294 |
| LMArena Creative Writing | 1104 | 1285 |
| LMArena Multi-Turn | 1111 | 1297 |
Frequently asked questions
Is DBRX better than DeepSeek-V2.5 (Sep 2024)?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 29.4 on the Noometry Index.
Is DBRX or DeepSeek-V2.5 (Sep 2024) better for coding?
DBRX scores higher on coding benchmarks: 32.9 versus 31.7 in the Noometry coding category.
How many benchmarks do DBRX and DeepSeek-V2.5 (Sep 2024) share?
19 benchmarks have published results for both models. DBRX has 21 scored results on Noometry and DeepSeek-V2.5 (Sep 2024) has 22.