🧠 Claude Sonnet 4 now supports 1M tokens and cut prices 🧠
5x increase enables processing entire codebases and dozens of research papers in single request

Anthropic has significantly expanded Claude Sonnet 4's capabilities with support for up to 1 million tokens of context, representing a 5x increase from previous limits. This breakthrough allows developers to process entire codebases with over 75,000 lines of code or analyze dozens of research papers in a single request.

The enhanced context window enables large-scale code analysis, document synthesis, and context-aware agents that maintain coherence across hundreds of tool calls. Long context support is now in public beta on the Anthropic API and Amazon Bedrock, with Google Cloud's Vertex AI coming soon.

🏆OpenAI beats 295/300 human programmers at IOI competition🏆
GPT-5 demonstrates continued AI progress despite industry slowdown concerns

OpenAI has achieved a remarkable milestone by beating all but 5 of 300 human programmers at the International Olympiad in Informatics (IOI), demonstrating that the company is not slowing down internally despite broader industry concerns about AI progress plateauing. This performance comes alongside GPT-5's impressive 74.9% score on SWE-bench Verified, a benchmark of real GitHub coding tasks, narrowly beating Anthropic's Claude Opus 4.1.

The results directly challenge the narrative that AI development is hitting diminishing returns, as OpenAI continues to achieve breakthrough performance metrics. GPT-5 is currently rolling out to Free, Plus, Pro, and Team users, with Enterprise and EDU access coming next week, as the company explores a $500 billion valuation.

📱 Apple's upcoming AI voice control will change iPhone usage 📱
New Siri features could be "under-the-radar hit" with iOS 26 development

Apple is developing new AI voice control features for Siri that could fundamentally change how people interact with iPhones. The upcoming voice control functionality is described as potentially being an "under-the-radar hit" if it works properly.

The development is connected to Apple's work on iOS 26 and "Liquid Glass 1.0," suggesting significant interface improvements ahead and the report indicates Apple is making substantial investments in voice-based AI interactions. This comes as Apple continues building its "Answers" team to develop a ChatGPT rival focused on world knowledge, while facing ongoing talent losses to competitors like Meta.

INTERESTING TO KNOW

🏆 Meta TRIBE model wins brain modeling competition🏆

Meta FAIR's Brain & AI team has won first place at the prestigious Algonauts 2025 brain modeling competition with their groundbreaking TRIBE (Trimodal Brain Encoder) model. The 1 billion parameter model represents the first deep neural network trained to predict brain responses to stimuli across multiple modalities, cortical areas, and individuals.

TRIBE combines pretrained representations from several Meta foundational models including Llama 3.2 for text, Wav2Vec2-BERT from Seamless for audio, and V-JEPA 2 for video processing. The system successfully predicted 80 hours per subject of spatio-temporal fMRI brain responses to movies from the Courtois NeuroMod project, marking a significant breakthrough in computational neuroscience and brain-AI interface research.

🏢 GitHub CEO steps down as Microsoft folds platform into CoreAI 🏢

GitHub CEO Thomas Dohmke has announced his resignation after nearly four years leading the platform, marking the end of GitHub's operational independence since Microsoft's $7.5 billion acquisition in 2018.

Microsoft is not replacing Dohmke's CEO position, instead folding GitHub's leadership team directly into its CoreAI organization led by former Meta executive Jay Parikh. Dohmke will remain through the end of 2025 to guide the transition before pursuing new startup opportunities. The move aligns GitHub more closely with Microsoft's AI platform strategy, as the CoreAI team focuses on building AI tools and platforms for both Microsoft and its customers.