AI Power Plays: Apple, OpenAI, and Mistral Make Their Moves
Date
June 11th, 2025
Reading Time
10 mins
Tags cloud
What's news
1. OpenAI’s open model is delayed
June 10, 2025
OpenAI's release of its first open model in years has been delayed until later this summer, after initially targeting an early summer launch. CEO Sam Altman announced the delay, citing unexpected and impressive progress from the research team that requires more time. The model is expected to match or exceed the reasoning capabilities of OpenAI’s o-series and outperform other open models like DeepSeek’s R1.
The open model's release comes amid growing competition from labs like Mistral and Qwen, which have recently launched advanced open reasoning models. OpenAI is also considering adding complex features, such as cloud integration to enhance its model's performance, though it's unclear if these will be included in the final release.
This launch is crucial for OpenAI’s reputation among researchers and developers, as the company has faced criticism for not open-sourcing its models. The open model is seen as a way to rebuild trust and compete in the increasingly crowded AI landscape.
Source: The Verge
2. Apple’s upgraded AI models underwhelm on performance
June 10, 2025
Apple has introduced two new AI models Apple On-Device and Apple Server. To power its Apple Intelligence features across iOS and macOS. However, Apple’s own benchmarks show these models still lag behind top competitors. The On-Device model performs comparably to similar sized models from Google and Alibaba, while the more powerful Server model was rated lower than OpenAI’s GPT-4o and even Meta’s Llama 4 Scout in both text and image tasks.
These results highlight Apple’s ongoing struggle to catch up in the fast moving AI race. Key projects like a revamped Siri remain delayed, and the company faces lawsuits over undelivered AI features. Still, Apple says both models show better efficiency and tool use, support 15 languages, and are now accessible to developers through its Foundation Models framework.
Source: TechCrunch
3. Mistral releases a pair of AI reasoning models
June 10, 2025
French AI lab Mistral has launched Magistral, its first family of reasoning focused models designed to tackle complex tasks like math, physics, and structured decision making. The lineup includes Magistral Small (24 billion parameters, open-source via Hugging Face) and Magistral Medium, which is available through Mistral’s Le Chat chatbot, its API, and partner cloud platforms.
While Magistral aims to support enterprise needs such as multi-step logic, risk modeling, and operational planning, early benchmarks show it lags behind competitors like Google’s Gemini 2.5 Pro and Anthropic’s Claude Opus 4 in key reasoning and programming tests. Despite this, Mistral promotes Magistral’s speed claiming 10x faster performance in Le Chat and wide language support including Arabic, Russian, and Chinese.
The release follows Mistral’s recent expansion into coding tools and enterprise AI, including its Le Chat Enterprise chatbot and Mistral Code client. Backed by over €1.1 billion in funding, Mistral continues to position itself as a serious contender in the competitive AI model space.
Source: Mistral AI
4. Apple lets developers tap into its offline AI models
June 9, 2025
At WWDC 2025, Apple introduced the Foundation Models framework, a new tool that lets developers use Apple’s AI models directly on device without needing cloud access. Part of the Apple Intelligence suite, this framework enables apps to deliver smarter, privacy-preserving features that work offline.
Apple says the framework supports Swift and is easy to implement developers can start with just three lines of code. It includes features like guided generation and tool calling. Early adopters include Automattic’s Day One app and AllTrails, which use it for journaling and personalized route suggestions. The framework is available now for developers, with a public beta coming next month.
Source: Apple
5. From spatial widgets to realistic Personas: All the visionOS updates Apple announced at WWDC
June 9, 2025
Apple's visionOS 26 update brings major improvements to the Vision Pro headset. New spatial widgets are customizable and blend into your environment, while the Photos app uses AI to add depth to 2D images. Safari supports immersive spatial browsing, and Personas for video calls are now more lifelike.
Users can now watch movies or collaborate in 3D with others, and enterprises get new APIs, device sharing, and privacy tools. The Logitech Muse accessory enables precise 3D input. Other updates include eye tracking scroll, iPhone unlock, and expanded language support.
Source: Apple
Tags cloud
Newsletter
DISCOVER MORE
![[Recap] UPP Global Technology JSC Establishing Anniversary](/homepage/news-section/new-4.webp)
LET’S TALK...
Content delivered to your inbox
ENTER YOUR EMAIL
YOU WANT TO...
Hanoi, Vietnam
Web3 Tower, No. 15, Alley 4, Duy Tan, Cau Giay, Hanoi, Vietnam