Mistral Large 4 enters preview; model weights are still to come
Mistral opens a preview of its new multimodal model. Access, document work and verifiable results are the useful questions for teams considering it.
Public preview announced on October 6.
Open model weights remain forthcoming.
Compare your own documents against known answers.
The short answer
A long report, an illustration and a specific question can quickly reveal whether an AI model helps with real work. That is where we would start with Mistral Large 4. Its preview offers a reason to compare another provider on the tasks your team actually handles.
Announcement date: 6 Oct 2026
What changed?
Mistral introduced Large 4 in public preview on October 6. Its documentation lists image understanding, document questions, structured outputs and tool calling. The changelog says open model weights are coming later.
Who is this relevant to?
Developers building document workflows or agents can assess whether the available service fits their needs. A future download would matter for operating the model independently; current plans should reflect the access available today.
What this means for your work
Our suggested comparison: choose an approved report containing a table and an illustration. Write down three questions and their correct answers before starting. Ask each model to identify the source location for every answer, then check it against the original. Record missing information, incorrect associations and manual corrections. This makes the comparison useful for a particular workflow rather than relying on an advertised ranking.
Limits of this report
This is an interpretation of provider documentation, not a hands-on test. Preview behaviour may change. Planned open weights are not an available download, and supported functions do not guarantee accurate answers.
Read next
AI research: from answer to reliable source
A compact workflow for checking claims transparently.
Introduce AI to a small team: run a pilot before adding subscriptions
Use real tasks, a simple evaluation sheet and a clear decision to find out whether an AI tool helps your team.
Check AI answers: turn source links into supported claims
A source link is not enough. Check what it actually supports and whether the context matches.
Was this article helpful?
Your feedback helps us improve our articles.
Votes are counted per language version. Repeat voting is limited; totals do not necessarily represent distinct people. Privacy