Mistral Small 4
Mistral Small 4 is a large language model made by Mistral, added to the catalog on 16 Mar 2026. It has a context window of 262,144 tokens and costs $0.15 per 1M input tokens and $0.60 per 1M output tokens. It accepts text and images and produces text. Typical uses: reasoning over long text inputs, working with text and image inputs and generating structured responses.
Price checked · source: OpenRouter
What Mistral Small 4 is
Mistral Small 4 is a multimodal model in the Mistral Small family. It accepts text and images as input and produces text. Its context window holds up to 262,144 tokens, and it can generate up to 209,715 tokens in a single output. This provides capacity for requests containing substantial source material and for extended text responses. The available description says the model unifies capabilities from several Mistral models, but does not name those capabilities or provide further detail about them. The listed input and output modalities are specific: the model takes text and image inputs, while its stated output is text.
Mistral Small 4 supports reasoning, tool use, structured outputs, and response-format controls. These features can suit tasks that require analysis, a specified response structure, or interaction with tools in a configured workflow. Image input allows visual material to be included with text in a request, while the model’s listed output remains text. Its large context window may be useful when a task involves long instructions or multiple pieces of source material. The catalog does not specify image-processing functions, particular tools, or supported response formats, so those details depend on the surrounding setup. Overall, Mistral Small 4 is suited to text-and-image requests, structured text generation, and reasoning-oriented work with substantial input context.
Key features
- Accepts text and image inputs and produces text
- Supports a context window of up to 262,144 tokens
- Can generate up to 209,715 tokens in one output
- Supports reasoning and tool use
- Supports structured outputs and response-format controls
Typical uses
- Reasoning over long text inputs
- Working with text and image inputs
- Generating structured responses
- Tool-assisted workflows
- Producing extended text outputs
Mistral Small 4 pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $0.15 |
| Output, per 1M tokens | $0.60 |
| Cached input read, per 1M tokens | $0.015 |
No price change since it was added on 16 Mar 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0003 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.0042 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.0057 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.023 |
Mistral Small 4 specifications
| Model id | mistralai/mistral-small-2603 |
|---|---|
| Version | mistral-small-2603 |
| Context window | 262,144 tokensAbout 196,608 English words |
| Max output | 209,715 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 16 Mar 2026 |
| Input types | text, image |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Mistral Small 4 yet.