Model comparison
Codestral vs Llama 13b
Codestral is the stronger model overall, scoring 30.6 to 24.4 on the Noometry Index.
Last verified . 0 shared benchmarks.
Mistral AI
MetaSummary
- The widest gap is in coding, where Codestral leads 27.3 to 21.4.
- Llama 13b has downloadable open weights; the other is API-only.
Side by side
| Codestral | Llama 13b | |
|---|---|---|
| Provider | Mistral AI | Meta |
| Noometry Index | 30.6 | 24.4 |
| Released | 2024-05-29 | 2023-02-24 |
| 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 Codestral leads
Codestral: 27.3 (#321), Llama 13b: 21.4 (#337)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| Aider Polyglot | 11.1% | — |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 683 |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Reasoning Codestral leads
Codestral: 19.8 (#251), Llama 13b: 14.0 (#329)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| LMArena Hard Prompts | — | 728 |
| BIG-Bench Hard | — | 37.9% |
| Epoch Capabilities Index | — | 100.58 |
| HellaSwag | — | 79.2% |
| LAMBADA | — | 75.2% |
| PIQA | — | 80.1% |
| WinoGrande | — | 73% |
Math Not comparable
Codestral: —, Llama 13b: 26.7 (#256)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| LMArena Math | — | 838 |
| GSM8K | — | 20.6% |
Knowledge Not comparable
Codestral: —, Llama 13b: —
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| ARC (AI2) Challenge | — | 52.7% |
| BoolQ | — | 78.7% |
| MMLU | — | 47.7% |
| OpenBookQA | — | 56.4% |
| TriviaQA | — | 77.9% |
Multimodal Not comparable
Codestral: —, Llama 13b: —
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual Not comparable
Codestral: —, Llama 13b: 16.6 (#297)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| LMArena Non-English | — | 819 |
Instruction Following Not comparable
Codestral: —, Llama 13b: 36.7 (#305)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| LMArena Instruction Following | — | 781 |
Writing & Preference Not comparable
Codestral: —, Llama 13b: 13.8 (#312)
| Benchmark | Codestral | Llama 13b |
|---|---|---|
| LMArena Text | — | 834 |
| LMArena Creative Writing | — | 794 |
| LMArena Multi-Turn | — | 753 |
Frequently asked questions
Is Codestral better than Llama 13b?
Codestral is the stronger model overall, scoring 30.6 to 24.4 on the Noometry Index.
Is Codestral or Llama 13b better for coding?
Codestral scores higher on coding benchmarks: 27.3 versus 21.4 in the Noometry coding category.
How many benchmarks do Codestral and Llama 13b share?
0 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Llama 13b has 21.