Best fit
Implementation and review tasks where Mistral models offer the right balance of speed, cost, and quality.
Reach for Mistral Vibe when the task wants Mistral models, cost-efficient passes, or a different reasoning style. ProPR keeps the same planning, isolated workspace, and GitHub PR boundary around the task. See all supported agents.
Implementation and review tasks where Mistral models offer the right balance of speed, cost, and quality.
Use the Mistral credentials configured in your self-hosted environment.
Mistral Vibe participates in the same plan, task, PR, follow-up, logging, and recovery model as other agents.
Vibe keeps the model list short. Both models run with a 256K context window, and the Vibe CLI also exposes a local option for self-hosted models. Model IDs work in issue labels, plan settings, /switch, and /review.
| Model | Context | Issue label | Reach for it when |
|---|---|---|---|
| Mistral Medium 3.5 | 256K | llm-vibe-mistral | The main implementation and review model — a cost-efficient default. |
| Devstral Small | 256K | llm-vibe-devstral | Light, fast coding passes where small and quick wins. |
Mistral Vibe runs on a Vibe subscription, a pay-as-you-go Mistral API key, or fully self-hosted open-weight models. You pay Mistral directly — or nothing, if you run the weights yourself. ProPR never marks up tokens.
| Plan | Price | Good for |
|---|---|---|
| Vibe Free | $0 | Trying Vibe with basic limits. |
| Vibe Pro | $14.99 / mo | More Vibe coding capacity for individuals. |
| Vibe Team | $24.99 / user / mo | Shared usage across a team. |
| Mistral API | Pay-as-you-go | Direct metered per-token billing. |
| Self-hosted | Free (your infra) | Run Devstral or other open-weight models yourself. |

ProPR runs the official Mistral Vibe CLI, and the setup specifics — CLI versions, model aliases, and credentials — are covered in the documentation. Inside ProPR, Mistral Vibe follows the same path as every other agent: planning, isolated implementation, pull request, follow-up, and recovery.