Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #14130

Mistral Large 4 (Le Chonk)

Related event timeline, sources and context from the news index.

EVENT TIMELINE

4

MODELS · 1 SOURCE · Mistral AI

Mistral’s Large 4 (Le Chonk) and Studio add managed agents, workflows and enterprise features

Mistral says its Mistral Large 4 model (Le Chonk) offers improved coding, cybersecurity and document reasoning for enterprise use, and that Mistral Studio now supports managed agents, workflows, apps, enterprise context, observability and MCP server integration for deployment across cloud, VPC or on‑prem environments. Customers including ABANCA, BNP Paribas, Accenture and Capgemini are cited as Studio users in production.

7.0

MODELS · 5 SOURCES · Mistral AI · WIRED AI · Hugging Face · The Decoder · TechCrunch AI

Mistral releases Mistral Large 4 (Le Chonk) weights — 1T-parameter open multimodal model

Mistral released the weights for Mistral Large 4 (Le Chonk), a 1‑trillion‑parameter natively multimodal model, claiming leading open‑weight results on coding and defensive cybersecurity benchmarks and new three-mode reasoning options to trade off cost and performance. The company says the model was trained on European infrastructure, is fully open‑weight, shows gains on FrontierCode (19.7%) and artificial analysis indexes (coding 48%, cyber 51%), and achieves a 96.6% average refusal rate on malicious cyber prompts while supporting self-hosted deployment.

8.0

COMPANIES · 1 SOURCE · Mistral AI

Mistral and NVIDIA partner to build AI-native engineering workflows and public physics benchmarks

Mistral and NVIDIA are collaborating to integrate Mistral Large 4 (Le Chonk) with NVIDIA accelerated computing (CUDA‑X, Warp) and Omniverse for GPU‑accelerated 3D simulation, aiming to speed simulations and enable closed‑loop AI-driven design. The partnership will produce public benchmarks and open datasets for physics models and agentic AI in CAE/EDA, expand NVIDIA‑powered infrastructure for Mistral, and builds on prior work including Mistral NeMo and the Nemotron Coalition; Mistral Large 4 (~1T total / 49B active params) was trained on ~3,800 Grace Blackwell GPUs and its weights are planned for release at the end of October.

8.0

MODELS · 3 SOURCES · TechCrunch AI · The Decoder · WIRED AI

Reflection unveils Beam, a 501B-parameter open-weight model claiming frontier performance at lower compute

Reflection AI announced Beam, a text-only, open-weight mixture-of-experts model with 501 billion total parameters (23 billion active), a 1 million token context window, and pretraining on 23.8 trillion tokens. The startup says Beam matches top Chinese open models on advanced reasoning benchmarks while using '3–4x less inference compute,' will publish weights and full technical details this month, and is targeting enterprises and sovereign 'AI factory' deployments; the performance claims have not been independently verified.

8.0