# This Week in All Things AI - Week 44-2025

> Sunday 26th October 2025 to Saturday 1st November 2025

**Published by:** [This Week in All Things AI](https://paragraph.com/@twiata/)
**Published on:** 2025-11-02
**URL:** https://paragraph.com/@twiata/this-week-in-all-things-ai-week-44-2025

## Content

MiniMax - M2 and Cursor 2.0 were what I what I term as this week big newsworthy items . The M2 team continues to work with various agentic coding harness to improve its integration which bodes very well for open source models I would like to encourage readers to write to me with their feedback about the content that I curate. Even short messages are appreciated Readers who think others in their family, friends and acquaintances who have a curiosity in knowing more about rapidly evolving AI tools/services/use cases would benefit from being subscribers of this weekly newsletter are encouraged to share this publication link to them and invite them to subscribe Below is my personal website which aggregates links to many of my socials as well as content and community that I curategoolamabbas | Instagram, TikTok | LinktreeChief Knowledge Officer @ Animoca Brands. Also assists in Growth Marketinghttps://linktr.eeThe following messages were posted on the 'All Things AI' Telegram group from Sunday 26th Oct 2025 to Saturday 1st Nov 2025Sunday 26th Oct 2025Monday 27th Oct 2025Chinese AI company MiniMax released its open-source MiniMax-M2 model , featuring a 230 billion parameter Mixture-of-Experts architecture that activates only 10 billion parameters for efficient performance on standard hardware. The model supports a 204,800-token context window and excels in benchmarks for mathematics, coding, and agentic reasoning, outperforming Claude 3 Opus in several areas while running at twice the speed of competitors and costing 8% of Claude 3.5 Sonnet per token. Free API access is available until November 7, with weights downloadable from Hugging Face and integration support from frameworks like vLLM and SGLang. Chinese AI company MiniMax released its open-source MiniMax-M2 model , featuring a 230 billion parameter Mixture-of-Experts architecture that activates only 10 billion parameters for efficient performance on standard hardware. The model supports a 204,800-token context window and excels in benchmarks for mathematics, coding, and agentic reasoning, outperforming Claude 3 Opus in several areas while running at twice the speed of competitors and costing 8% of Claude 3.5 Sonnet per token. Free API access is available until November 7, with weights downloadable from Hugging Face and integration support from frameworks like vLLM and SGLang. https://x.com/MiniMax__AI/status/1982674798649160175 OpenRouter @OpenRouterAI Free new model from MiniMax w/ top-of-the-market intelligence for open source Try it out! Demis Bellot @demisbellot New MiniMax M2 (A10B/230B MoE) is free on @OpenRouterAI 267 11:25 AM • Oct 27, 2025 Yifan Zhang @yifan_zhang_ Some fun facts about Minimax M2: 1. Minimax uses GPT-OSS-like structure, i.e., Full Attention interleaved with Sliding Window Attention (SWA). 2. It uses QK Norm, and every single attention head has its own unique, learnable RMSNorm. 3. The full attention and SWA parts Skyler Miao @SkylerMiao7 MiniMax M2: Our advanced model that achieves global Top 5 status, surpassing Claude Opus 4.1 and ranking just behind Sonnet 4.5. Coming October 27th - Stay Tuned 585 12:34 PM • Oct 27, 2025 Tuesday 28th Oct 20251Poster: Deus Machine We recently launched our developer copilot for the Ethereum network and lots of people who used it have good things to say. You can access it at ethereum.bytebell.ai. https://www.reddit.com/r/ethereum/comments/1of9qyf/i_built_an_ai_that_actually_knows_ethereums/ It reached 100000 views in Two days, And developers have asked more than 1500+ queries in 2 days # Bytebell: Verified answers for engineering, not vibes ## Problem Engineering teams bleed $47K per dev per year searching across repos, docs, PDFs, tickets, and Slack. Even billion parameter models hallucinate without enterprise context and cannot cite where answers came from. That's the purpose behind Bytebell. Please reach out to me if you are interested in this solution2Qualcomm unveils two AI inference chips, the AI200, set for 2026, and the AI250, planned for 2027, and says Humain is the first customer https://www.cnbc.com/2025/10/27/qualcomm-ai200-ai250-ai-chips-nvidia-amd.html Tareq Amin @TareqAmin_ Thrilled to announce the next major step in our partnership with @Qualcomm: the launch of the world’s first fully optimized edge-to-cloud AI infrastructure in Saudi Arabia. With 200MW of AI200 and AI250 rack solutions, the partnership will enable faster, more efficient 298 11:41 PM • Oct 27, 2025 https://www.qualcomm.com/news/releases/2025/10/humain-and-qualcomm-to-deploy-ai-infrastructure-in-saudi-arabia- Wednesday 29th Oct 20251Cartesia raises $100M and launches Sonic-3 state-of-the-art model for realtime conversation and built on SSM (State Space Models) instead of Transformers Karan Goel @krandiash We've raised $100M from Kleiner Perkins, Index Ventures, Lightspeed, and NVIDIA. Today we're introducing Sonic-3 - the state-of-the-art model for realtime conversation. What makes Sonic-3 great: - Breakthrough naturalness - laughter and full emotional range - Lightning fast - 8,246 12:00 AM • Oct 29, 2025 https://x.com/i/trending/1983222742062072180 The below newsletter has a very deep dive on Mamba, which is a SSM https://newsletter.maartengrootendorst.com/p/a-visual-guide-to-mamba-and-state2Starting Nov 4, users in India can enjoy ChatGPT Go free for an entire year, no strings attached! * No card required * Existing users also get 12 months free * ChatGPT Go (₹399/month) → Now ₹0 for a full year https://techcrunch.com/2025/10/27/openai-offers-free-chatgpt-go-for-one-year-to-all-users-in-india/Thursday 30th Oct 20251Cursor Unveils 2.0 with In-House Composer AI Model for Rapid Agentic Coding Cursor launched version 2.0 of its AI-powered code editor, introducing the proprietary Composer-1 model, an in-house mixture-of-experts system trained via reinforcement learning on coding tasks. The update enables multi-agent workflows with up to eight parallel agents for planning, editing, and testing, alongside features like a built-in browser, sandboxed terminals, voice mode, and hybrid model integration for enhanced development efficiency Cursor @cursor_ai Introducing Cursor 2.0. Our first coding model and the best way to code with agents. 7,843 12:12 AM • Oct 30, 2025 Sasha Rush @srush_nlp Composer is a new model we built at Cursor. We used RL to train a big MoE model to be really good at real-world coding, and also very fast. cursor.com/blog/composer Excited for the potential of building specialized models to help in critical domains. 757 12:32 AM • Oct 30, 2025 Cognition (makers of Devin and acquirer of Windsurf) releases SWE-1.5, their fast agent model which is available in Windsurf Cognition @cognition Today we’re releasing SWE-1.5, our fast agent model. It achieves near-SOTA coding performance while setting a new standard for speed. Now available in @windsurf. 945 6:30 AM • Oct 30, 2025 https://cognition.ai/blog/swe-1-52via Google --- So excited and thankful for our partnership with Reliance we've just start rolling out FREE Google AI Pro plans across India for eligible Jio customers! (total value ₹35,100) Grab it in the MyJio app or here: https://www.jio.com/google-gemini-offerFriday 31st Oct 2025Saturday 1st Nov 2025I welcome feedback not only on this newsletter but across all the various surfaces where I curate content The cover image of this newsletter via generated via the Seedream 4.0 model from Bytedance within the Freepik tool via the following prompta sacred mountain pass in Chinese ink painting style, mist flowing between jagged peaks rendered with delicate brushstrokes, a monk ascending step by step with prayer beads swaying in rhythm, robes whispering in the wind, minimal palette of black ink and soft washes of gray, atmosphere infused with solitude and devotion, emotion of reverence and quiet endurance

## Publication Information

- [This Week in All Things AI](https://paragraph.com/@twiata/): Publication homepage
- [All Posts](https://paragraph.com/@twiata/): More posts from this publication
- [RSS Feed](https://api.paragraph.com/blogs/rss/@twiata): Subscribe to updates
- [Twitter](https://twitter.com/yusufg): Follow on Twitter