Magistral Medium
Mistral AI's Magistral reasoning model via the Vercel AI Gateway: native chain-of-thought reasoning with vision and function calling at a 128k context window. Reasoning is on by default. Premium tier. Uses 2 premium requests per send before length multipliers. Supports image input. Does not support web search.
You always get the exact model you pick — we never silently route you to another.
About Magistral Medium
Magistral Medium is Mistral's reasoning model — its answer to chain-of-thought systems — and on just4o.chat it runs through the Vercel AI Gateway as mistral/magistral-medium. Unlike the rest of the Mistral lineup here, reasoning is on by default: the model works through problems step by step before answering, which helps on math, logic, multi-step analysis, and structured decision-making. It also accepts image input, so visual reasoning is on the table, and it supports function calling for tool use. The context window is 128k tokens — smaller than the 256k on Mistral Large 3 and Medium 3.5, but ample for most reasoning tasks. It is a Premium model on just4o.chat, billed at 2 premium requests per send before length multipliers, at a provider price of $2.00 per million input and $5.00 per million output tokens. Two caveats: no native web search, and because it reasons before answering, responses can take longer and consume more output tokens than a non-reasoning model. For hard problems where a careful, worked answer beats a fast one, that trade is usually worth it.
Best for
- Multi-step reasoning — math, logic puzzles, and structured problem-solving
- Analytical tasks that benefit from explicit step-by-step thinking
- Visual reasoning over images, charts, and diagrams
- Tool-using agents that need a reasoning model with function calling
- Decision-support and planning tasks where answer quality matters more than latency
Specifications
| Provider | Mistral |
|---|---|
| Released | 2025-06 |
| Intelligence | High |
| Speed | Medium |
| Context window | 128,000 tokens |
| Knowledge cutoff | Not officially published |
| Input price | $2.00 / 1M tokens |
| Output price | $5.00 / 1M tokens |
| Request cost | 2 premium requests |
| Plan tier | Premium |
| Input | Text, Image |
| Output | Text |
| Features | Native reasoning (chain-of-thought on by default), 2 premium requests per send before length multipliers, Function calling supported, Image input supported, Served through the Vercel AI Gateway (mistral/magistral-medium), Mistral applies zero data retention to gateway traffic |
| Model ID | magistral-medium |
Frequently asked questions
It is a reasoning model: chain-of-thought is on by default, so it works through problems before answering. The other Mistral models on just4o.chat are general or coding models that answer directly.
Through the Vercel AI Gateway's OpenAI-compatible endpoint under the mistral/magistral-medium route. Mistral applies zero data retention to gateway traffic.
Yes — Magistral Medium accepts text and image input, so it can reason over visual material.
$2.00 per million input tokens and $5.00 per million output tokens at provider reference pricing. On just4o.chat it is a Premium model that uses 2 premium requests per send before length multipliers.
128,000 tokens — smaller than Mistral Large 3 and Medium 3.5, but sufficient for most reasoning tasks.