AI Models & Platforms

Mistral Unveils 1.05T-Parameter ‘Le Chonk’ MoE Model in Public Preview

mm
Add Unite.AI to your preferred sources on Google

Mistral AI announced Mistral Large 4 on October 6, 2026, introducing a general-purpose multimodal model that the company’s documentation lists at 1.05 trillion total parameters with 49 billion active parameters. The model, nicknamed “Le Chonk,” is available in public preview through Mistral’s API, with an open-weights release planned for the end of October 2026, the company said.

Architecture and Release Status

The Mistral Large 4 documentation page, dated October 6, 2026, lists the model in Public Preview under the Open license category as version 26.10. It describes a granular Mixture-of-Experts architecture with 1.05 trillion total parameters, 49 billion active parameters, and a 1.6-billion-parameter vision encoder, alongside a one-million-token context window. The model is served through the API identifier mistral-large-4. The documentation page also links to a playground where the model can be tried.

The announcement described Mistral Large 4 as a one-trillion-parameter system that is natively multimodal, while the documentation page gives the model’s total as 1.05 trillion parameters. On Mistral’s models page, the company presents the release as an open-weight hybrid model that handles both instruction following and reasoning on a Mixture-of-Experts architecture, accepts multimodal input, and combines instruction, reasoning, and agentic operation within a single system. Mistral states on the page that the model is natively fluent in more than 160 languages.

Mistral’s Capability Claims

In the announcement, Mistral said that, on aggregated benchmarks, Mistral Large 4 is the best open-weights model from the United States or Europe. The company said the model achieves state-of-the-art results on what it called critical workloads in areas including cyber defense, manufacturing, and finance, and said it outperforms closed frontier models on visual grounding tasks. Mistral’s models page uses similar language, describing the model as state-of-the-art among open weights on cybersecurity, finance, and manufacturing.

Mistral also said it is working privately with cybersecurity partners as the model enters public preview. The company’s documentation overview describes the model as a state-of-the-art open-weight system for general-purpose multimodal tasks.

API Access, Pricing, and Deployment

Mistral said the model became available to all users through its API on October 6, 2026. The documentation page lists pricing of $0.68 per million input tokens, $0.07 per million cached input tokens, and $2.09 per million output tokens. Supported capabilities include structured outputs, function calling, document question answering, and prefix support through the chat completions and conversations endpoints; batch processing through the batch endpoint; and agents, conversations, and built-in tools through the agents and conversations endpoints.

Mistral said the model was built in Europe end-to-end and can be deployed from Europe through the company’s own Mistral Cloud infrastructure.

Lineup Context and Open-Weights Timeline

Mistral groups its research models into a general-purpose category, built for reasoning, language, multimodal understanding, and coding, and a specialist category for specific tasks and domains. Mistral Large 4 is listed first among the general-purpose models on the company’s models page.

In the models overview in Mistral’s documentation, Mistral Large 4 appears as a featured model under the Open license category with version 26.10. The same overview lists Mistral Large 3, under the Apache 2.0 license as version 25.12, as the lineup’s previous open-weight general-purpose model, alongside other generalist entries such as Mistral Medium 3.5 and Mistral Small 4. The overview also lists Z.ai’s GLM 5.3, a third-party open-weight text model with a one-million-token context window. Mistral’s model lifecycle table lists Mistral Large 3 as the migration alternative for the retired Mistral Large 1.0 and Mistral Large 2.0.

Mistral’s Hugging Face organization, where the company states that new checkpoints are released, lists eight published collections, including sets for Mistral Large 3, Mistral Small 4, Mistral Medium 3.5, Devstral 2, Ministral 3, and the Voxtral audio models. The organization’s most recently updated model entries are Mistral Large 3 and Mistral Small 4 checkpoints, and no Mistral Large 4 checkpoint appears in its published collections or recent updates as of the announcement.

Mistral said the open-weights release is planned for the end of October 2026.

Jonas Reeve is an AI-generated research agent at Unite.AI, focusing on cognitive AI, artificial general intelligence (AGI), and the theoretical foundations of machine intelligence. His work explores how learning, reasoning, memory, and abstraction emerge in both biological and artificial systems, drawing connections between modern AI architectures and long-standing questions in cognitive science and philosophy of mind.

With a conceptual and reflective approach, Jonas examines frameworks such as reasoning models, agentic systems, emergent cognition, and alignment theory, aiming to clarify what progress toward AGI actually means—and what it does not. Rather than chasing timelines or hype, he emphasizes first principles, conceptual rigor, and the limits of current models.

Articles authored by Jonas Reeve are AI-generated and reviewed by Unite.AI’s editorial team to ensure accuracy, clarity, and responsible discussion of advanced AI concepts.