Mistral Medium 3.5 is one of the more interesting model launches of April 2026 because it is being sold in more than one context at once. There is the clean API story, where pricing is easy to read. Then there is the product story, where Medium 3.5 powers Le Chat and Mistral Vibe, and the cost picture becomes broader than simple token math.
That distinction matters. Teams searching for “Mistral Medium 3.5 pricing” are often mixing together three different questions: what the API costs, what Mistral Pro or Team costs, and whether remote coding-agent usage is fully included or only included up to a point.
Mistral Medium 3.5 API pricing
Mistral lists Medium 3.5 at $1.50 per 1 million input tokens and $7.50 per 1 million output tokens. The model card positions it as a frontier-class multimodal model optimized for coding and agentic work, with a 256k context window.
Those rates make the model easier to understand than many rivals. There are no layered tiers on the base API card itself. If you are consuming Medium 3.5 through the API, the budgeting logic is mostly about how much input context you keep attached and how verbose the model becomes on output-heavy runs.
| Mistral Medium 3.5 API item | Official price |
|---|---|
| Input | $1.50 / 1M tokens |
| Output | $7.50 / 1M tokens |
| Context window | 256k |
That output price is the part to watch. Medium 3.5 can look affordable in evaluation because short tests are input-dominant. On real coding and research workflows, output often becomes the larger budget driver.
How Medium 3.5 fits into Le Chat and Mistral Vibe pricing
Mistral’s consumer and team pricing adds another layer. On Mistral’s pricing page, Pro is listed at $14.99 per month and Team at $24.99 per user per month, while Enterprise is custom.
The key line for developers is that Pro includes Mistral Vibe for all-day coding, PAYG beyond. That wording is important. It suggests that some coding usage is bundled, but not that every heavy agent workflow is unlimited at the subscription price.
| Mistral plan | Listed price | What matters for Medium 3.5 buyers |
|---|---|---|
| Pro | $14.99/month | Includes Le Chat upgrades and Mistral Vibe for all-day coding, with pay-as-you-go beyond included usage |
| Team | $24.99/user/month | Adds collaborative workspace features for team usage |
| Enterprise | Custom | Used for larger deployments, controls, and custom commercial arrangements |
If you only look at the subscription headline, Medium 3.5 can seem dramatically cheaper than the API. But that is only true for users whose behavior stays inside the included product limits. Once usage spills into pay-as-you-go territory, the API economics start mattering again.
Why bills can still rise even if the model price looks simple
The biggest pricing trap with Medium 3.5 is assuming that a simple rate card means a simple bill.
First, the model is built for long-horizon work. Mistral says it was designed for multi-tool, structured-output, coding, and agent workflows, and it also powers remote agents in Vibe. That means tasks can run longer, produce more output, and touch more systems than a normal chat workflow.
Second, product usage and API usage are not the same thing. A developer may start inside Pro, then move into team collaboration, heavier Vibe use, or direct API integration. At that point, the cost center shifts from seat pricing to token pricing.
Third, output-heavy coding sessions are where the economics move fastest. Planning, patch generation, test analysis, summaries, retries, and tool-oriented explanations all create output tokens. Medium 3.5’s output is five times more expensive than its input, so verbose workflows are where teams most often underestimate spend.
When Mistral Medium 3.5 pricing makes the most sense
Choose Pro or Team first if:
- your main use case is human-driven coding inside Le Chat or Vibe
- you want a low-friction way to try the model in real work
- you value a seat-based starting point before moving to custom infrastructure
Choose the API first if:
- you are building your own product or agent backend
- you need cost accounting at the workflow level
- you want direct control over prompt design, tool use, and token efficiency
- you expect usage to exceed what bundled product plans comfortably cover
For many businesses, the smartest path is mixed: let humans explore the workflow with Pro or Team, then move repeatable production tasks to the API once usage patterns are clear.
What teams should budget for Mistral Medium 3.5 in 2026
A practical budgeting approach is to think in layers.
Evaluation layer: use Pro if you are testing Medium 3.5 as a daily coding assistant or agent surface.
Team adoption layer: move to Team when collaboration, domain verification, and managed workspace features start to matter.
Production layer: budget directly against the API rates once you are embedding Medium 3.5 into software, remote agents, or repeatable business workflows.
The broader takeaway is that Medium 3.5 pricing is simple only at the model card level. At the product level, it is really a blended pricing story: subscriptions get you in the door, but serious agentic usage still pushes you back toward pay-as-you-go economics.
That is what teams should budget around, not just the headline seat price.