Model comparison
Codestral vs DBRX
Codestral is the stronger model overall, scoring 30.6 to 29.4 on the Noometry Index.
Last verified . 2 shared benchmarks.
Mistral AI
DatabricksSummary
- They share 2 benchmarks with published results for both. Codestral scores higher in 0 categories and DBRX in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where DBRX leads 32.9 to 27.3.
- DBRX has downloadable open weights; the other is API-only.
Side by side
| Codestral | DBRX | |
|---|---|---|
| Provider | Mistral AI | Databricks |
| Noometry Index | 30.6 | 29.4 |
| Released | 2024-05-29 | 2024-03-27 |
| Weights | Proprietary | Open |
| Context window | 256K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.30 | — |
| Output $ / M tokens | $0.90 | — |
| Results tracked | 7 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DBRX leads
Codestral: 27.3 (#321), DBRX: 32.9 (#266)
| Benchmark | Codestral | DBRX |
|---|---|---|
| HumanEval+ | 73.8% | 70.1% |
| MBPP+ | 61.9% | 55.8% |
| Aider Polyglot | 11.1% | — |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1132 |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
Reasoning DBRX leads
Codestral: 19.8 (#251), DBRX: 21.4 (#222)
| Benchmark | Codestral | DBRX |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| LMArena Hard Prompts | — | 1113 |
Math Not comparable
Codestral: —, DBRX: 24.3 (#269)
| Benchmark | Codestral | DBRX |
|---|---|---|
| LMArena Math | — | 1145 |
| MATH Level 5 | — | 11.7% |
Knowledge Not comparable
Codestral: —, DBRX: 14.9 (#294)
| Benchmark | Codestral | DBRX |
|---|---|---|
| GPQA Diamond | — | 32.9% |
| LMArena Expert | — | 1076 |
Multilingual Not comparable
Codestral: —, DBRX: 29.3 (#268)
| Benchmark | Codestral | DBRX |
|---|---|---|
| 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
Codestral: —, DBRX: 57.5 (#270)
| Benchmark | Codestral | DBRX |
|---|---|---|
| LMArena Instruction Following | — | 1112 |
Long Context Not comparable
Codestral: —, DBRX: 33.7 (#258)
| Benchmark | Codestral | DBRX |
|---|---|---|
| LMArena Longer Query | — | 1112 |
Writing & Preference Not comparable
Codestral: —, DBRX: 33.6 (#275)
| Benchmark | Codestral | DBRX |
|---|---|---|
| LMArena Text | — | 1119 |
| LMArena Creative Writing | — | 1104 |
| LMArena Multi-Turn | — | 1111 |
Frequently asked questions
Is Codestral better than DBRX?
Codestral is the stronger model overall, scoring 30.6 to 29.4 on the Noometry Index.
Is Codestral or DBRX better for coding?
DBRX scores higher on coding benchmarks: 32.9 versus 27.3 in the Noometry coding category.
How many benchmarks do Codestral and DBRX share?
2 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and DBRX has 21.