Model comparison
Codestral vs Falcon-40B
Codestral has enough public results to be ranked (#290); Falcon-40B does not yet, so treat this comparison as directional.
Last verified . 0 shared benchmarks.
Mistral AI
Technology Innovation InstituteSummary
- Falcon-40B has downloadable open weights; the other is API-only.
Side by side
| Codestral | Falcon-40B | |
|---|---|---|
| Provider | Mistral AI | Technology Innovation Institute |
| Noometry Index | 30.6 | — |
| Released | 2024-05-29 | 2023-03-15 |
| Weights | Proprietary | Open |
| Context window | 256K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.30 | — |
| Output $ / M tokens | $0.90 | — |
| Results tracked | 7 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Codestral: 27.3 (#321), Falcon-40B: —
| Benchmark | Codestral | Falcon-40B |
|---|---|---|
| Aider Polyglot | 11.1% | — |
| BigCodeBench Instruct | 41.8% | — |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Reasoning Not comparable
Codestral: 19.8 (#251), Falcon-40B: —
| Benchmark | Codestral | Falcon-40B |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| BIG-Bench Hard | — | 37.1% |
| Epoch Capabilities Index | — | 104.45 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 77.3% |
| PIQA | — | 83% |
| WinoGrande | — | 76.9% |
Math Not comparable
Codestral: —, Falcon-40B: —
| Benchmark | Codestral | Falcon-40B |
|---|---|---|
| GSM8K | — | 33.8% |
Knowledge Not comparable
Codestral: —, Falcon-40B: —
| Benchmark | Codestral | Falcon-40B |
|---|---|---|
| ARC (AI2) Challenge | — | 61.9% |
| BoolQ | — | 83.1% |
| MMLU | — | 56.9% |
| OpenBookQA | — | 56.6% |
| TriviaQA | — | 79.9% |
Frequently asked questions
Is Codestral better than Falcon-40B?
Codestral has enough public results to be ranked (#290); Falcon-40B does not yet, so treat this comparison as directional.
How many benchmarks do Codestral and Falcon-40B share?
0 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Falcon-40B has 12.