● Covered by 5 sources · 5 reportsMedium impact3 neutral2 positive
Mistral Releases Weights for Multimodal Mistral Large 4, Nicknamed "Le Chonk"
Key points
- Mistral released weights for its Mistral Large 4 (ML4) model, "Le Chonk."
- ML4 is a multimodal, multilingual MoE model with 1T total parameters.
- It accepts text and images as input, generating text output.
- The model shows improved performance in coding and cybersecurity tasks.
- ML4 is distributed under the Mistral Ecosystem Open-Weight License.
Mistral Large 4 Weights Released
Mistral has released the weights for its Mistral Large 4 (ML4) model, nicknamed "Le Chonk." This release follows a public preview of the model earlier this month. The weights are available on Hugging Face, allowing businesses and governments to download and deploy the system independently.
ML4 is a multimodal, multilingual Mixture-of-Experts (MoE) model. It is designed to take both text and images as input and generate text. The model has approximately 1 trillion total parameters, 49 billion active parameters, and a context window of 1 million tokens.
Performance and Capabilities
Mistral Large 4 has shown improved performance in key areas, particularly agentic coding and defensive cybersecurity. On the Terminal-Bench 4.0 coding benchmark, ML4's score increased to 41% from its preview version's 28%, placing it alongside models like Codex GPT-6.1 Terra. It also scores 82% on the Artificial Analysis Cyber Index test.
The model is described as being particularly strong for cybersecurity tasks, coding agents, and multimodal applications. It combines a large pool of experts with sparse attention and a Pixtral vision encoder, selecting a small group of experts for each token and specific context positions.
Open-Weight Distribution and Licensing
The weights for Mistral Large 4 are distributed under the Mistral Ecosystem Open-Weight License, which is an MIT derivative with limitations for model-as-a-service offerings. This open-weight approach allows for customizable, self-hosted deployment options, providing organizations with greater control over advanced AI capabilities.
Mistral emphasizes the importance of open weights for business continuity and to ensure AI acts as a tool for empowerment rather than dominance. The company states that ML4 achieves state-of-the-art performance at a fraction of the compute and capital committed by closed labs.
Impact on the AI Landscape
The release of ML4's weights signifies a continued trend of open models becoming increasingly competitive with proprietary models. Mistral aims to set a new standard for Western open models and intends to continue advancing the frontier of open models in agentic coding and cybersecurity.
The model is available in two official variants, with GGUFs also created by Unsloth and the llama.cpp team. This release is positioned to empower sovereign AI applications by enabling independent deployment and customization.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
GitHub huggingface/blog GitHub huggingface/trl GitHub NVIDIA-NeMo/Automodel GitHub NVIDIA/TensorRT-LLMHow outlets covered it
Mistral has released the weights for its trillion-parameter open-source AI model, Mistral Large 4 (ML4), enabling businesses and governments to deploy it independently. This move provides organizations with greater control over advanced AI capabilities, particularly for cybersecurity applications, and positions ML4 as a more economical alternative to existing open-weight models.
Mistral has released the open weights for its Large 4 (ML4) model, which it states has improved significantly since its preview three weeks ago. The company claims ML4 is the leading open agentic coding model, outperforming GLM-5.3 in cybersecurity and coding benchmarks.
Mistral has released the weights for Mistral Large 4 (Le Chonk), a 1-trillion-parameter multimodal model, following a public preview. The model shows improved performance in coding and defensive cybersecurity, and offers three reasoning modes for cost-performance balance. This release allows for customizable, self-hosted deployment options and is positioned for sovereign AI applications.
Mistral has released the weights for its Mistral Large 4 (ML4) model, nicknamed 'Le Chonk,' which is a trillion-parameter model. ML4's coding performance on Terminal-Bench 4.0 improved to 41%, placing it alongside models like Codex GPT-6.1 Terra and demonstrating competitive performance against other open and proprietary models.
Mistral has released Mistral Large 4, a multimodal, multilingual Mixture-of-Experts (MoE) model, on Hugging Face. This model accepts text and images as input to generate text, featuring 1T total parameters, 49B active parameters, and a 1M token context window. It is particularly effective for cybersecurity tasks, coding agents, and multimodal applications, and is distributed under the Mistral Ecosystem Open-Weight License.