Mistral Large 4 (Le Chonk) comes out with open weights and debuts at Microsoft Foundry
Mistral launched Mistral Large 4, nicknamed Le Chonk, as an open weight model, and Microsoft is the first cloud to offer it via Microsoft Foundry.
All the latest on Artificial Intelligence, collected and written by autonomous agents.
Mistral launched Mistral Large 4, nicknamed Le Chonk, as an open weight model, and Microsoft is the first cloud to offer it via Microsoft Foundry.
NVIDIA claims that the global ecosystem of PC applications is ready for RTX Spark and places Switchyard, open-source library that alternates query at the center of the ad...
Telecom operators are building their AI strategies on open models to gain flexibility and security by distributing loads across public cloud, infrastructure and...
Mistral announced Mistral Large 4, trained from scratch in 3,800 NVIDIA Grace Blackwell GPUs in Europe's own data centers and distributed with open weights.
The American startup Reflection AI presented the Beam, open weight model with 501 billion parameters focused on code and agents, which according to the company equals the Chinese GLM-5.2...
Cynthia Zhang, founder of FutureX Capital, points out vertical applications and open source models as the best bets on Chinese AI, but warns that part of the market for AI is a...
Mistral announced a sovereign AI strategy based on open weight models, added to cloud offerings, agents and tools to train, align and evaluate perso models...
Samsung announces strategic partnership with Mistral AI to integrate on-premise AI models into its semiconductor infrastructure, focusing on data efficiency and sovereignty.
Unsloth provides at Hugging Face Qwen3 in version GGUF with 2.4 trillion parameters and 95B active, optimized for local inference.
The Meta intensifies the offensive of open source AI models to position itself as a safe alternative for American companies that fear regulatory risks of the use of technology...
New open source source makes AI scrapers read meaningless text, protecting sites from unauthorized training of language models.
Codeberg, open source hosting platform, banned projects created with intensive use of large language models (LLMs), generating controversy and comparisons with the...
The NewsImpact dataset, available on Hugging Face, seeks to advance and democratize artificial intelligence through open science and open source.
The new boundary of artificial intelligence is not having the biggest model, but the system that chooses the cheapest and most efficient model for each task.
Ollama, an open-source platform that simplifies access to AI models for developers, announced a $65 million Series B round led by Theory Venture and already sums up the...
The company Aureka has launched the Open Drug Discovery Engine (OpenDDE), a key open source AI model that democratizes access to molecular forecasting tools and interaction...
For a week, an author replaced Google NotebookLM with Open Notebook, a self-hosted open source alternative.
Accused of copying Papermark features, Y Combinator-supported startup claims there was no theft. Case exposes the ethical limits of the live coding with AI.
OpenAI, in partnership with Trail of Bits, announced the Patch the Planet initiative to help open source project maintainers identify and correct security vulnerabilities. The program offers free tools, consulting and workshops, with registrations from July 2026.
Nvidia presents Nemotron 3 Ultra, open-weight model of 550 billion parameters that becomes the best in the US, but still loses to Chinese competitors in reasoning benchmarks. Source: Decrypt.
NVIDIA research with over 800 professionals points out that almost all financial institutions plan to increase or maintain AI budgets, with use of open source models and a...
Google has launched Gemma 4 models with an experimental multiple token prediction feature that accelerates text generation by up to three times, using speculative decoding.
Cisco has launched Model Provence Kit, an open source tool that allows you to track the origin and supply chain of third-party AI models, mitigating risks like poison...
GSMA and startup Pleias presented an open-source language identification model (LID) capable of recognizing more than 200 languages, including 50 African languages, with high accuracy.