Mistral AI has launched Le Chat Enterprise, powered by its new Mistral Medium 3 model. The platform includes enterprise search, agent builders, custom data connectors, document libraries, and hybrid deployment options, with all features rolling out over the next two weeks.
Mistral AI has released Mistral OCR, an API for extracting text and images from documents at $1 per 1,000 pages (approximately $0.50 with batch inference). The company claims 94.89% overall accuracy on its internal test set, comparing favorably to GPT-4o (89.77%), Gemini 2.0 Flash (88.69%), and Azure OCR (89.52%).
Mistral AI has released a comprehensive update to le Chat, introducing Flash Answers feature with ~1000 words/sec generation speed, iOS and Android apps, and new Pro ($14.99/month) and Team subscription tiers. The company also announced Enterprise tier in private preview with custom model deployment options.
Mistral AI announced a partnership with Agence France-Presse to integrate AFP's newswire content into Le Chat. The assistant will access 2,300 daily stories across six languages—French, English, Spanish, Portuguese, German, and Arabic—from AFP's network of 1,700 journalists.
Mistral AI has released major updates to le Chat, its free AI assistant, adding web search with citations, a canvas interface for collaborative editing, and image generation powered by Black Forest Labs' Flux Pro. The updates are powered by the new Pixtral Large multimodal model and are available free in beta.
Zhipu AI released GLM-5.2, a 753 billion parameter model with a 1 million token context window. The model scores 62.1% on SWE-bench Pro and introduces IndexShare architecture that reduces per-token FLOPs by 2.9× at 1M context length. Released under MIT license with no regional restrictions.
Google has released Gemini 3.1 Flash Image, internally codenamed "Nano Banana 2," an image generation and editing model with a 131K context window. The model is priced at $0.50 per 1M input tokens and $3 per 1M output tokens.
Google has released Nano Banana Pro, an advanced image generation and editing model built on Gemini 3 Pro. The model supports 2K/4K output resolution, preserves identity across up to five subjects, and includes real-time Search grounding for context-rich visual synthesis.
Midjourney announced its first hardware product: a full-body ultrasound scanner using 40 Butterfly Network imaging modules and 2 petaflops of processing power. The company plans to install 10 units in a San Francisco spa location opening before the end of 2027.
Google is rolling out Android Bubbles support for the Gemini overlay, allowing users to minimize active conversations into a floating spark logo icon. The feature prevents chat loss when switching away from the overlay, similar to Gemini Live's floating waveform.
Cohere has released North Mini Code, the first model in its North family and its first agentic coding model. The sparse mixture-of-experts architecture features 30B total parameters with 3B active, a 256K-token context window, and up to 64K tokens of output, available free via OpenRouter under Apache 2.0 license.
Amazon has launched autonomous agents in QuickSight (branded as Quick) that execute tasks continuously in the background while users attend meetings or focus on other work. The update includes 16 new data source integrations, an activity feed that consolidates communications across tools, and cross-system query capabilities that join data from multiple sources in real time.
Epic Games released Unreal Engine 5.8 with an experimental Model Context Protocol plugin that allows developers to connect AI models including Claude and Gemini to the game engine. The company plans to make MCP integral to Unreal Engine 6, expected in late 2027.
GitHub has updated Copilot's architecture to optimize token consumption through improved context handling and model routing. The changes aim to make user credits last longer by reducing unnecessary token usage in coding sessions.
GitHub has improved how Copilot handles context and routes requests to models, reducing token usage per session. The changes aim to make user credits last longer by eliminating wasted tokens.
OpenAI has launched scheduled tasks in ChatGPT, allowing users to automate reminders, recurring work, and monitoring. The feature is rolling out today to Plus, Pro, Business, and Enterprise users, and will replace the existing Pulse feature in 14 days.
Replit is now available as a direct integration within Claude, according to Replit. The integration allows users to design applications in Claude Design using natural language, then send projects directly to Replit for development and deployment without manual copy-pasting or context switching.
Google is shutting down Gemini CLI and Gemini Code Assist for GitHub for consumer-level users on June 18, 2026, as it consolidates development tools into its Antigravity CLI platform. Enterprise users will retain access to Gemini Code Assist unchanged.
Amazon Web Services has released new capabilities for Bedrock AgentCore, its platform for building AI agents. The update includes a managed knowledge base that handles vector storage and retrieval across enterprise data sources, native web search using Amazon's Alexa infrastructure, and a payment system enabling agents to access paid content and APIs.
Google Vids now offers AI avatars to free personal accounts, providing 10 video generations per month that can be split between avatars and Veo video generation. The Workspace video creation tool has reached 7 million monthly active users.