Mistral 7B
Mistral AI's first model, a 7.3B dense model under Apache 2.0, beats Llama 2 13B on all benchmarks the company reports.
A 7.3B model that Mistral says matches a Llama 2 model more than 3x its size on some metrics, with no usage restrictions. The first release from a new lab, and a template for small, permissively licensed models.
- Date
- Wednesday, 27 September 2023
- Lab
- Mistral AI
- Kind
- open-weights
- Access
- open weights
Figures
| Measure | Value | Measured by |
|---|---|---|
| Parameters | 7.3B dense, Apache 2.0 | company |
Comparison to Llama 2 13B is company-reported.
Sources
This record was checked against its sources on 6 October 2026. How we check