Mistral Large 4 enters public preview as trillion-parameter MoE
10/09/2026 — 10/11, 01:16·2 sources (all)·2 reports (all)
Key takeaways
- Mistral has moved its new flagship, Mistral Large 4, into public preview. The model is a trillion-parameter MoE with native text and image s…
Background & analysis
On October 9, 2026, according to a DEV Community article, Mistral moved its new flagship model, Mistral Large 4, into public preview. The model is described as a trillion-parameter, multimodal mixture-of-experts system with native support for text and images, in which roughly 49 billion to 52 billion parameters are active per token. The article notes that Mistral Large 4, also referred to as ML4, carries the playful nickname "le Chonk."
The same report says the model is positioned for complex work such as cybersecurity, finance, and agentic reasoning. Beyond its scale, the article highlights Mistral's commitment to release the model weights by the end of the month, continuing the company's push toward open, self-hostable foundation models.
As it stands, the model is in public preview and the weights have not yet been released. The only timing given so far is Mistral's statement that they will arrive within the month.
AI-generated from 1 reports · updated 1 day ago
Latest turnAccording to a DEV Community post, Mistral has moved its new flagship Mistral Large 4 into public preview: a trillion-parameter, natively multimodal MoE model with roughly 49–52B parameters active per token. The post says it targets cybersecurity, finance and agentic reasoning, and that Mistral has committed to releasing the weights within the month.
- Heat index
- 86
- All sources
- 2
- All reports
- 2
- First seen
- 1 day ago
Reports on this story headlines open the original
According to a DEV Community post, Mistral has moved its new flagship Mistral Large 4 into public preview: a trillion-parameter, natively multimodal MoE model with roughly 49–52B parameters active per token. The post says it targets cybersecurity, finance and agentic reasoning, and that Mistral has committed to releasing the weights within the month.
DEV Community · AIAI score 61Tom's Hardware reports that Artificial Analysis scored Mistral's newly released Large 4 at 38, placing it behind open models from Xiaomi, Z.ai, Moonshot, and DeepSeek. The number comes from a third-party benchmarking platform and reflects where the model sits on public leaderboards.
Tom's HardwareAI score 53
Other stories people are talking about
- Heat index 597OpenAI Defends Firing of Three Safety Researchers, Denies Safety Concerns15 sources
- Heat index 438TypeSafe AI raises $870M led by a16z at $7.5B valuation8 sources
- Heat index 433OpenAI in Early Talks for $30B Raise at $1.4T Valuation12 sources
- Heat index 307Microsoft launches Decision-1 for structured decisions5 sources
- Heat index 256Anthropic Tightens Claude Usage Rules on Elections and Weapons7 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion