Mistral Large 4 Listed: 524k Context, Multimodal Input at $0.68/$2.09

What Mistral Large 4 Is

Mistral has listed Large 4, an open-source multimodal model that accepts text and image inputs and returns text. The context window sits at 524k tokens — long enough to swallow large codebases, lengthy document sets, or extended conversation histories without chunking.

Pricing

Input is priced at $0.68 per million tokens; output at $2.09 per million tokens. The input price is competitive for a model of this claimed capability tier — comfortably below several closed-source large-model alternatives. Output pricing is mid-range: not the cheapest you'll find, but not aggressive either. If your workload is read-heavy (long context in, short answers out), the economics look reasonable. Generation-heavy pipelines — summarization at scale, long-form drafting — will want to model that $2.09 output rate carefully before committing.

Open-Source Matters Here

Being open-source means you're not locked into Mistral's API. Teams with the infrastructure to self-host can drive per-token costs down significantly, making the listed API price more of a ceiling than a fixed cost.

Who Should Pay Attention

• RAG and document-analysis teams — 524k context reduces retrieval complexity for many use cases.
• Multimodal app developers — image-plus-text input without switching models is a workflow simplification.
• Cost-sensitive API users — $0.68 input undercuts several comparable-tier models listed on our site.

As always, benchmark on your own data before switching. Context window size and modality support only matter if the output quality holds up for your specific task.

See Mistral Large 4 — live specs & price history for up-to-date pricing and head-to-head comparisons.

Ezra, Scout AI Team