Mistral Large 4

Mistral's 1T-parameter multimodal MoE flagship; open weights published on Hugging Face under a modified-MIT license.

Mistral AI · Frontier/foundation models · model

Published in Feed

Overview

Mistral AI released Mistral Large 4, nicknamed "Le Chonk," as a public preview API on October 6, 2026, and has since published open weights on Hugging Face as Mistral-Large-4-1T-A52B plus an NVFP4 checkpoint for Blackwell GPUs. It is a mixture-of-experts model of about 1 trillion parameters (about 52B active, 128 routed experts with 4 active per token plus 1 shared expert) with a 1M-token context window, reading text and images and answering in text across 160+ languages. The weights ship under the Mistral Ecosystem Open-Weight License, a modified MIT license that allows commercial use but requires a separate commercial license to run a Model-as-a-Service business. Per the model card, it scores 26 on the Artificial Analysis Intelligence Index v5, which the card describes as the highest among Western open-weight models.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • About 1T total / 52B active parameters, 128 routed experts (4 active) plus 1 shared expert
  • Open weights on Hugging Face: Mistral-Large-4-1T-A52B and an NVFP4 checkpoint for Blackwell GPUs
  • Mistral Ecosystem Open-Weight License (modified MIT); MaaS businesses need a separate commercial license
  • 1M-token context; text and image input, text output; 160+ languages
  • Reasoning effort selectable: low, medium or high (default high)
  • Served via vLLM (vllm/vllm-openai:mistral-large-4, TP 8) and SGLang images; trained on 3,800 NVIDIA Grace Blackwell GPUs

Use cases

  • Multilingual enterprise deployments spanning EU languages
  • Document and image understanding tasks requiring text output
  • Sovereign or on-premises self-hosting on 8-GPU nodes
  • Agentic coding workloads (Artificial Analysis Coding Agent Index 48%, per model card)

Pricing

API: $1.36/M input, $4.18/M output tokens (list; promotional $0.68/$2.09 reported at launch). Open weights free to download under the Mistral Ecosystem Open-Weight License; hosting it as a paid MaaS requires a commercial license.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
6 Oct 2026
Record updated
30 Oct 2026
Record last checked
30 Oct 2026

Explore the Feed

Other entries filed under frontier/foundation models; inclusion does not imply equivalent features.

  • Muse Wonder

    Meta's most capable Muse models, with active memory and long-goal focus for agents, 1M context and $4/$20 per million tokens.

  • Rho-1

    Reka's 19B omni-reasoning model that handles text, images, video and robot actions in one network, in research preview.

  • Amazon Nova 3

    Amazon's third-generation Nova frontier model family, unveiled at an AWS keynote with a 1M-token context window and a dedicated agentic tool-use variant.

Browse the Feed