Mistral Large 4

Mistral Large 4 is Mistral's flagship, announced on 6 October 2026 as a public API preview. It is a mixture-of-experts model with about 1.05T total and 52B active parameters, a 1.6B vision encoder and a 1M-token context, combining instruct and reasoning modes. Mistral says the weights will be released openly later in October 2026.

By Mistral AIReleased October 6, 2026Data checked October 10, 2026

PreviewClosedFrontier

Specifications

Context window: 1Mabout 750,000 words, or 1,500 pages
Max output: Not published
Input price: $1.36per 1M tokens
Output price: $4.18per 1M tokens
Cached input: $0.14per 1M tokens
Weights: Closed
Parameters: 1.05T total, 52B activeMoE
Input types: Text, Images
Output types: Text
Reasoning mode: Yes
Tool calling: Yes
Released: Oct 6, 2026Preview

Mistral list price per Mistral docs. During the public preview the docs show a sale price of $0.68 input / $0.07 cached / $2.09 output. API-only for now: Mistral promises open weights 'by the end of the month' (October 2026), but none are on Hugging Face yet and no license is named, so weights is 'closed' and license is null. Context 1M per Mistral docs.

What it costs in practice

Workload	Monthly cost at list price
Support chatbot30,000 conversations a month, 2,000 input and 400 output tokens each	$132
Document summaries2,000 documents a month, 12,000 input and 800 output tokens each	$39
Coding agent1,000 tasks a month, 60,000 input and 6,000 output tokens each	$107

Standard rates, no caching or batch discounts. Run your own numbers.

Strengths

1M-token context window
Hybrid instruct and reasoning model with image input
Mistral-operated European deployment option

Trade-offs to weigh

Preview status; weights and license not yet published
Self-hosting a ~1T-parameter model needs multi-GPU servers

Where it fits

Mistral AI positions Mistral Large 4 for:

agentsreasoninglong documentsmultimodalenterprise

Where you can use it

Alternatives to consider

Model	Tier	Context	Input / output per 1M
Grok 4.7xAI	Frontier	500K	$2 / $6	Compare
Qwen3.8-MaxAlibaba (Qwen)	Frontier	1M	$2 / $6	Compare
GLM-5.3Z.ai (Zhipu AI)	Frontier	1M	$1.40 / $4.40	Compare
DeepSeek-V4-ProDeepSeek	Frontier	1M	$1.32 / $3.96	Compare

Frequently asked questions

How much does Mistral Large 4 cost?

Mistral Large 4 lists at $1.36 per million input tokens and $4.18 per million output tokens, with cached input at $0.14 (checked October 10, 2026). Mistral list price per Mistral docs. During the public preview the docs show a sale price of $0.68 input / $0.07 cached / $2.09 output. API-only for now: Mistral promises open weights 'by the end of the month' (October 2026), but none are on Hugging Face yet and no license is named, so weights is 'closed' and license is null. Context 1M per Mistral docs. Confirm current rates on the provider's pricing page before you commit.

What is the context window of Mistral Large 4?

Mistral Large 4 accepts up to 1,000,000 tokens of context (about 750,000 words, or 1,500 pages). Long prompts cost more and can be slower, so most apps send only what the task needs.

Is Mistral Large 4 open source?

No. Mistral Large 4 is a proprietary model from Mistral AI; its weights aren't published. You can use it through the Mistral AI API and through the provider's apps.

What is Mistral Large 4 best for?

Mistral AI positions Mistral Large 4 for agents, reasoning, long documents, multimodal. It accepts text, images as input, has a reasoning (thinking) mode and supports tool calling. Test it on a sample of your own tasks before you choose it for production.

Sources

Facts checked against these pages on October 10, 2026. Spot an error? Tell us.

Still choosing a model?

Send us your use case and constraints. We'll come back with a shortlist and an evaluation plan you can run on your own data.

Free and vendor-neutral: no model provider pays for placement.
Your details go only to partners you agree to.